Dr. Rameshwar K. Sharma, DAB No. 1431 (1993)

  Department of Health and Human Services

        Departmental Appeals Board

  RESEARCH INTEGRITY ADJUDICATIONS PANEL

SUBJECT:        Dr. Rameshwar K. Sharma    DATE:  August 6, 1993 Docket
   No. A-93-50 Decision No. 1431

   DECISION

Dr. Rameshwar K. Sharma (Respondent) requested a hearing before the
Research Integrity Adjudications Panel to contest two findings of
scientific misconduct by the Office of Research Integrity (ORI) of the
Public Health Service (PHS), as well as the administrative actions which
PHS proposed to take against him as a result.  The findings of
scientific misconduct involve statements in two grant applications
submitted by Respondent to PHS which ORI determined to constitute
falsifications.  In each case, ORI found that Respondent purported to
have data which he did not in fact possess.  The Respondent admitted
that the statements were erroneous, but asserted that they were not
intentional misstatements and did not constitute scientific misconduct.

On the basis of its findings, ORI proposed to (1) prohibit Respondent
from serving on PHS advisory committees, boards, or peer review groups
for three years and (2) require any application to PHS by Respondent
during that period to contain a certification by the applicant
institution that "the application has been reviewed and approved by all
investigators involved in the project," or by "at least two qualified
scientists," if no other scientists are involved in the project.

     SUMMARY OF DECISION

After providing factual and procedural background information, our
decision addresses the standards of conduct which ORI must prove were
violated in order to establish scientific misconduct.  Since there were
no standards explicitly defined in applicable regulations or law during
the time in question, ORI was obligated to show by evidence what
standards were applicable to similarly-situated researchers.  We find
that ORI proved the existence of standards applicable during the period
in question that would be violated, at a minimum, by intentional,
material falsification of research methods and results in a grant
application.  We find that ORI did not prove that negligent inclusion of
a false statement in a grant application in the circumstances involved
in this case would so seriously deviate from accepted practice at the
time as to constitute scientific misconduct.  Thus, ORI could establish
scientific misconduct in this case only if it proved by a preponderance
of the evidence that Dr. Sharma intentionally falsified material
statements in his grant applications.

ORI found scientific misconduct because it concluded that Dr. Sharma
intentionally included erroneous statements in two grant applications
submitted to NIH.

In regard to the first application, ORI inferred that Dr. Sharma acted
intentionally based on its position that Dr. Sharma had offered
inconsistent explanations of how the error occurred.  ORI also argued
that Dr. Sharma's guilty intent was shown by (1) his failure to place a
citation after the sentence at issue, (2) a pattern of misstatements,
(3) his failure to report the errors to NIH upon discovery, and (4) his
general lack of credibility.  We find that ORI failed to prove by a
preponderance of the evidence that there was intentional deception in
the first grant application.

It is important, when examining Dr. Sharma's allegedly inconsistent
statements, to keep in mind as background the nature of the error
involved.  The sole error -- the core of this dispute which has consumed
so much time and attention in PHS and elsewhere -- is a subscript error
in a single sentence.  Subscripts are important in distinguishing one
protein from another in the same family.  The subscript error in
question, however, is inconspicuously located and forms no part of any
obvious web of deceit.  The application contains none of the data or
explanations which reviewers of the applications would have expected to
see if the erroneous statement were true.  Consequently, the error was
unlikely to mislead reviewers -- and, in fact, the evidence does not
establish that it did.  On the other hand, the record does show,
virtually without challenge, that a typographical error could have
occurred as a result of one wrong keystroke using word-processing
macros.  (The two subscripts which were interchanged in the error in
question appeared about 128 times in that one application.)  The record
also shows that the same subscript transposition by Dr. Sharma and
others occurred many times in circumstances where it was much more
obviously an inadvertent error.  All this had to be kept in mind when
testing ORI's case, relying as it does upon collateral actions and
statements of Dr. Sharma which ORI alleged cast the significance of
intentional deceit upon an otherwise unremarkable error.

We have analyzed in detail the statements made by Dr. Sharma which ORI
interpreted as inconsistent explanations for the misstatement in the
first grant application.  Although some of his statements are ambiguous,
and at times reflect apparent communication problems, none of his
statements, individually or collectively, persuasively demonstrates any
intent to deceive.  We therefore decline to draw the inference from
these statements, suggested by ORI, that Dr. Sharma acted intentionally.
We reject ORI's other arguments because--

 (1) ORI did not establish that an additional citation was
 required after the sentence at issue because, but for the error,
 the reference would have been clear from the context;

 (2) ORI did not prove a pattern of false statements in this
 application or other writings by Dr. Sharma;

 (3) ORI did not prove that Dr. Sharma had a duty to report to
 NIH errors discovered only after the grant applications had been
 rejected for funding and an investigation was already underway;
 and

 (4) ORI's attacks on Dr. Sharma's credibility were not supported
 by the record.

In regard to the second grant application, ORI relied heavily on an
inference that intentional deception in this grant application was part
of the pattern of misstatements about the progress of his research as
evidenced by the first grant application.  We reject this inference for
two reasons.  First, it is dependent on intentional deception in the
first application, which ORI failed to establish.  More importantly, the
second grant application was hastily and poorly written (as Dr. Sharma
admitted) and contained numerous statements directly contradicting the
erroneous statement.

Therefore, we find that ORI failed to carry its burden of proving
scientific misconduct by a preponderance of the evidence as to either
grant application.  Consequently, we conclude that ORI's findings are
not supported and the proposed administrative actions are not justified.

         Background

Factual and Scientific Background 1/

During the time that the two applications at issue were prepared and
submitted, Dr. Sharma was a researcher in the Department of Brain and
Vascular Research at the Cleveland Clinic Foundation (CCF).  Two of the
major research areas in his laboratory involved --

 (1) investigating the structure and function of à2 adrenergic
 receptors, and in particular an à2 receptor which was believed
 to be unusual in that it coupled with guanylate cyclase and
 which was therefore assigned the name à2GC (the subject of
 Exhibit H-2, which was a grant application submitted by Dr.
 Sharma to NIH [National Institutes of Health] in February 1989,
 but not funded) 2/, and

 (2) investigating the role of an enzyme, identified as a 180 kD
 protein, in carrying signals across cell membranes and the
 mechanisms for its regulation (the subject of Exhibit H-3, which
 was a grant application submitted by Dr. Sharma to NIH in June
 1989, but not funded).

See Transcript of Hearing, held May 13-19, 1993 (Tr.), at 869; Exs. H-2
and H-3.

Both research projects involved study of the structure of  particular
proteins (the à2GC receptor protein or the 180 kD enzyme protein).
Proteins are molecules which are constructed in linear chains of amino
acids and which are vital to the structure and functioning of living
cells.  Tr. at 200, 518; ORI Scientific Terms at 1.  Receptors are large
proteins on the surface of cells which interact with molecules in the
outside environment.  Tr. at 517-9; ORI Scientific Terms at 3.  Various
kinds of receptor proteins can be defined by their biochemical activity,
for example, adrenergic receptors bind to epinephrine molecules.  Tr. at
519; ORI Scientific Terms at 4; Exhibit H-38, at 315-8.  (The main
distinguishing feature of the particular adrenergic receptor which Dr.
Sharma was researching was that it was believed to couple to guanylate
cyclase, rather than adenylate cyclase, which was more common.  Tr. at
520, 522-3.)

Proteins can be identified by their structure, in addition to their
biochemical behavior.  One way to research the structure of a protein is
to seek to identify the specific sequence of amino acids which form the
protein.  Tr. at 202-3.  An initial approach is to cut the protein into
fragments, called peptides, using enzymes or chemicals which reliably
divide the chain at predictable points.  Tr. at 523.  These fragments
can be isolated and analyzed to create "peptide maps," which show the
pattern of pieces from these breaks as a unique "fingerprint."  Tr. at
491-3, 524-5; ORI Scientific Terms at 1.  A peptide map does not
disclose the precise order of the amino acids in a protein, although it
may point to areas of difference between similar proteins.  Tr. at 493,
506-7; ORI Scientific Terms at 2.  Additional work must be done to
determine the amino acid sequence of a particular peptide, and then to
determine the order of the sequence within the protein as a whole.  Tr.
at 527-8.

Each protein is produced by a particular gene, which is composed of DNA
(deoxyribonucleic acid) nucleotide base sequences.  Genes do not
directly produce the protein and often contain additional material which
is not directly translated in the creation of protein.  Tr. at 541,
948-9.  Rather, the DNA in a gene is translated into the shorter form of
RNA (ribonucleic acid, also composed of nucleotides), and the RNA in
turn translates its sequence into corresponding amino acids to produce
the protein.  Tr. at 541-2, 948.  Each of the twenty common amino acids
can be coded for by one or more sets of three nucleotides, known as
"codons."  Tr. at 200; 528-9.

Once an adequate portion of the sequence of amino acids in a protein is
known, a possible next step is to create a probe for molecular cloning.
A probe is a piece of cDNA ("complementary DNA") which is synthesized
from nucleotide codons that correspond to the known amino acid sequence.
Tr. at 206-7, 527-9.  Such a probe, known as an oligomer probe, can be
used to screen a "library," i.e., a collection of copies of all the RNA
(reverse translated into cDNA form) in a particular cell type.  Tr. at
540, 949-50.  Sometimes, more than one codon triplet can code for a
particular amino acid, a phenomenon known as the "degeneracy" of the
code.  Tr. at 228-9, 528-30, 928.  In such situations, the scientist
constructing a probe has several alternatives.  The scientist can
construct a series of probes embodying each of the possible
alternatives, known as "degenerate probes."  Tr. at 531-2.  The
scientist can select a region of the sequence to construct the probe
which has the least degree of degeneracy.  Tr. at 229-30; 530-1.  The
scientist can construct a probe by guessing the most likely combination
of codons, known as a "guessmer" probe.  Tr. at 532-3, 928.  In making a
guessmer probe, the choice of likely codons is assisted by the use of
codon bias data tables which set out the tendency in particular species
to prefer certain codon alternatives over others in coding for
particular amino acids.  Tr. at 530.  The probe constructed by the
scientist is then introduced into the library under specified conditions
("stringency"), and the results are studied to see whether the probe has
reacted with, or "hybridized to," any cDNA in the library.  Tr. at
542-6, 589-9.  The isolated cDNA to which the probe has hybridized is
known as a cDNA clone.  Tr. at 547.  The cDNA may contain all or only a
part of the sequences necessary to produce the protein.  Tr. at 214,
550.

In the absence of data on the sequence of amino acids in a desired
protein, scientists can also use probes based on related genes to screen
for molecular clones of the desired protein that will hybridize to the
probe as well.  Tr. at 208, 927; see generally, Exhibit H-38, at
104-112.  This approach is feasible because hybridization can occur with
cDNA of other, closely-related proteins.  Id.  In fact, unintended
mismatches occur at times, either because the clone is not identical to
the probe, but contains some matching sequences, or because the wrong
codons were selected.  See Tr. at 533, 598, 608, 627-30.  Related
proteins, or "families" of proteins, often have areas within their amino
acid sequences which are identical, or "conserved," among the family
members.  Tr. at 534-5.  If a probe is based on a conserved portion of
the amino acid sequence of one family member, it would, under
appropriate conditions, have a good chance of hybridizing to other
members of the same family.  Tr. at 537-8.  Another strategy would be to
create a probe that was based on the entire gene sequence of one of the
family members for which cDNA had already been cloned.  Tr. at 538-9.
Such a probe would be much larger than an oligomer probe (about 200-1000
nucleotides, as opposed to 15-50 nucleotides).  Tr. at 539.  In order to
identify the cDNA which was cloned by a probe, a scientist would need to
(1) use the resultant clone to produce, or "express," the protein and
identify the protein by its chemistry and behavior, or (2) sequence the
clone and compare it to the protein's amino acid sequence if known.  Tr.
at 548-50.  This step would be necessary, even if the probe was based on
sequence data from the desired protein, rather than a related one,
because of the possibility of mismatches or conserved sequences
hybridizing to something other than the target cDNA.  Tr. at 955-6.

It is uncontested that Dr. Sharma did not have the amino acid sequence
data of either of the proteins that he was studying at the time he
submitted the grant applications at issue, although he had made efforts
to obtain such data.  In the case of à2GC, another laboratory published
amino acid sequence data on a related member of the same family of
proteins, i.e., à2A, derived from human platelet sources.  Dr. Sharma
used probes based on amino acid sequence data of à2A to screen a library
from rat adrenocortical carcinoma cells in an effort to clone à2GC.  His
application sought funding to determine if the clones which he had
obtained did, in fact, represent the unique receptor, which he had named
à2GC.  In the case of 180 kD, Dr. Sharma had peptide maps but did not
have amino acid sequence data.  In that application, he sought funding
to continue further work on sequencing, characterizing and studying the
functioning of the protein.

Procedural History

The initial inquiry in this case was triggered on April 20, 1990, when
Stephen Chalberg, Ph.D., a researcher working under Dr. Sharma,
approached Carlos M. Ferrario, Ph.D., the chairman of Dr. Sharma's
department, with concerns that certain grant applications submitted by
Dr. Sharma contained inaccurate statements.  ORI noted that these
allegations arose "in an atmosphere of conflict" between Dr. Ferrario
and Dr. Sharma over various issues of personnel, job performance and
funding.  Exhibit R-10 (ORI Final Report) at 4.  In addition, Dr.
Chalberg was already anxious at that time to leave Dr. Sharma's
laboratory.  Tr. at 148.  Dr. Ferrario referred the matter to Dr.
Bernadine P. Healy, M.D., then Chair of the Research Institute at CCF,
who convened a preliminary inquiry committee (Healy Committee), which
found no evidence of scientific misconduct, as opposed to sloppiness, in
the allegations before it (which did not include those relating to
Exhibit H-2).  Ex. H-4; Ex. H-5, at 5.  A second inquiry committee,
chaired by Thomas A. Hamilton, Ph.D. (Hamilton Committee), was convened
(for reasons which are disputed) and concluded that the allegations
against Dr. Sharma had substance.  Ex. H-7, at 4.  Therefore, a formal
investigation was instituted, chaired by Andrew J. Fishleder, M.D.
(Fishleder Committee).  The Fishleder Committee concluded that Dr.
Sharma should be "exonerated from the allegation of scientific
misconduct" and found no "conclusive evidence" of an intent to mislead
the study sections reviewing the grant applications to make
recommendations to NIH about funding.  Ex. H-8, at 3-4.

The matter then passed from CCF to the PHS, which began its
investigations in November of 1990.  After finding CCF's procedures
inadequate, the Office of Scientific Integrity (OSI) undertook its own
investigation, which was interrupted when the case was transferred to
the Office of Scientific Integrity Review (OSIR) in October 1991.  ORI
Final Report at 7.  Subsequently, the functions of OSI and OSIR were
transferred to ORI.  The ORI investigation involved interviews of
relevant witnesses, a site visit, and consultation with scientific
advisors.  Id. at 9-11.  In April 1992, OSIR issued draft reports
finding no scientific misconduct.  Exs. R-5 and R-6.  After receiving
comments on the draft reports, obtaining additional submissions, and
conducting a second interview with Dr. Sharma, ORI issued its Final
Report and Action Notice finding misconduct in regard to the allegations
set forth below.

Dr. Sharma appealed to the Research Adjudications Integrity Panel on
December 21, 1992.  After preliminary proceedings, a hearing was held in
the case, May 13-19, 1993, followed by post-hearing briefing by both
parties.

    Issues

ORI found that Dr. Sharma made false statements in two grant
applications, submitted to PHS, in order to mislead reviewers into
believing his research to be more advanced than it was in fact.  Tr. at
11-24.  The specific allegations at issue were set forth in ORI's Final
Report as follows: 3/

 [Allegation 1:]  ORI finds that Dr. Sharma falsified statements
 on pages 21, 27 and 32 of Application 1 [Exhibit H-2 in this
 proceeding], and, consequently, Dr. Sharma committed scientific
 misconduct.  Specifically, ORI finds that Dr. Sharma falsely
 reported the results of research conducted in his laboratory by
 stating that the probe used to screen the library and isolate
 the cDNA clones was based on the amino acid sequence data of the
 à2GC receptor when the work was done using the à2A receptor
 because Dr. Sharma did not have the amino acid sequence of the
 à2GC receptor.

   *   *   *

 [Allegation 4:]  The ORI finds that the statement appearing on
 page 23, Application 3 [Exhibit H-3], was falsified.
 Specifically, ORI finds that Dr. Sharma falsely claimed to
 possess unpublished sequence data of the 180-kD protein.

ORI Final Report at 54 and 55.

ORI contended that intentionally claiming in a PHS grant application to
have data which an applicant does not possess in order to mislead
reviewers constituted scientific misconduct because it violated accepted
standards of conduct in the scientific community during the relevant
period.  Dr. Sharma admitted that he did not have amino acid sequence
data on either the à2GC receptor or the 180 kD protein at the time these
grant applications were filed.  However, he contended that the
statements which appeared to claim such possession were not
intentionally misleading, but were caused by errors.  Respondent
Post-Hearing Br. at 1-2, 5-6.

Thus, the key issue to be decided is whether ORI proved by a
preponderance of the evidence on the record before us that the
misstatements in each grant application were the result of intentional
deception.

ORI also argued in the alternative that Dr. Sharma was guilty of
scientific misconduct even if the statements were not intentional, if he
should have known about the false statements and was negligent in
failing to prevent their inclusion in the grant applications.  ORI
contended that Dr. Sharma failed to follow accepted scientific
practices, such as requesting colleagues to preview his grant
applications, including more cites, or reporting the errors to NIH upon
discovery.  Thus, we must also decide whether ORI established that
negligent inclusion of a false statement in a grant application
constituted scientific misconduct under standards applicable at the
time.  We therefore turn next to defining the applicable legal
standards.

        Applicable Legal Standards

The role of the Panel is to review all the evidence before it to reach a
de novo decision as to the existence of scientific misconduct and the
appropriateness of the proposed actions, rather than to review ORI's
procedures or the reasonableness of ORI's decisions on the evidence
before ORI.  Ruling on Respondent's Motion to Dismiss the Complaint,
dated May 10, 1993 (Ruling) at 10, n.8. 4/  The Panel's decision herein
"will be final agency action" on the proposed administrative actions in
this matter. Guidelines on Hearings Before the Research Integrity
Adjudications Panel (Guidelines) at 5.

The issues in this matter were bifurcated during a preliminary
conference, held on January 19, 1993, and the parties briefed legal
issues raised by Respondent in advance of the hearing held to resolve
factual disputes.   In a motion to dismiss, Respondent (1) challenged
the authority of PHS to take administrative action against scientific
misconduct committed prior to the issuance of regulations at 45 Fed.
Reg. 32,446 (August 8, 1989) (1989 Regulations) defining such
misconduct, (2) asserted that PHS discretion did not extend to the
actions proposed here and was not supported by a notice announcing such
actions in cases of scientific misconduct because the notice at 56 Fed.
Reg. 27,384 (June 13, 1991) violated the Administrative Procedure Act, 5
U.S.C. . 553, and (3) argued that PHS's authority in any case did not
reach misrepresentations in grant applications which were never funded.

The Panel denied Respondent's motion to dismiss, concluding that PHS had
authority to propose administrative actions against scientific
misconduct committed before the 1989 Regulations in order to protect the
integrity of federal research grant funds.  The Panel rejected
Respondent's arguments concerning the Administrative Procedure Act.
Furthermore, the Panel concluded that "[m]aterial falsifications in
grant applications relating to the principal investigator's research
capacity and accomplishments are clearly among the types of conduct
which PHS may investigate and propose actions for under its
discretionary authority."  Ruling at 2.  However, the Panel also
concluded that no general policy had been established specifying the
applicable standards of conduct or identifying particular administrative
actions appropriate to specific instances of scientific misconduct.
Therefore, a "case-by-case adjudication" is required to determine:  "1)
what conduct a researcher has engaged in; 2) whether that conduct
violated applicable standards existing at the time, derived either from
the scientific community or from federal requirements for applying for,
conducting, or reporting federally supported scientific research; and 3)
whether the particular administrative actions proposed are appropriate."
Ruling at 3.  The Panel further noted that, under the Panel review
process adopted by PHS, PHS has the burden of proving scientific
misconduct by a preponderance of the evidence, which encompasses "both
the existence of the conduct and the standard which applies, including
whether Respondent had whatever level of intent is required to violate
that standard."  Ruling at 3, 13; Guidelines at 6.

We found that no law or regulations established a specific definition of
scientific misconduct against which to measure Dr. Sharma's conduct.
However, we recognize here, as we have in other proceedings, that the
definition adopted by PHS in 1989 acts as a limit on the scope of these
proceedings.  The 1989 Regulations define scientific misconduct as --

 fabrication, falsification, plagiarism, or other practices that
 seriously deviate from those that are commonly accepted within
 the scientific community for proposing, conducting or reporting
 research.  It does not include honest error or honest
 differences in interpretations or judgments of data.

54 Fed. Reg. at 32,449.  Thus, ORI must establish the applicable
standards on a case-by-case basis and must exclude "honest error," for
example, or any conduct which does not "seriously deviate" from accepted
practices.

ORI established at the hearing by testimony from scientists with
experience as grant applicants and reviewers covering the relevant field
at the time in question that intentional falsification of a material
element of a grant application would be a serious deviation from
accepted practices.  See, e.g., Tr. at 211 (Dr. Marshak agreeing it is
not "appropriate" in the scientific community to represent in a grant
application that one has data which one does not yet have), 470 (Dr.
Merrick agreeing it is "an improper practice to submit a grant
applications that includes data that you do not possess"), 678-82.  Dr.
Sharma himself indicated that he thought a wrong statement in a grant
application should be reported promptly to NIH, because his philosophy,
for his students as well as himself, was "to have the utmost integrity"
and not to put in data which is merely anticipated.  Tr. at 847-8 (Dr.
Sharma).  ORI thus could establish scientific misconduct if it proved by
a preponderance of the evidence that Dr. Sharma intentionally falsified
material statements in his grant applications.

In its post-hearing brief, ORI argued for the first time that the
requisite legal standard for intent was conclusively resolved by an
earlier debarment case, and that it included a negligence standard.  ORI
Post-Hearing Br. at 5-7, citing Robert Edward McCaa, Ph.D., DAB No. 823
(1987).  On that basis, ORI asserted that, in order to "satisfy the
requisite level of intent," ORI had only to "prove that Dr. Sharma . . .
should have known that the false statements were in the applications."
ORI pointed to a statement in that case that the researcher, Dr. McCaa,
"knew or should have known" that he did not have sufficient information
to accurately report the methods or results of experiments set forth in
a published paper.  McCaa at 33.

After reading the McCaa decision as a whole, we conclude that ORI's
reliance on that decision is clearly misplaced, for the following
reasons:

o       The statement that Dr. McCaa "knew or should have known" of
inaccuracies was clearly intended as a factual finding regarding the
differing states of Dr. McCaa's knowledge of various inaccuracies in the
published paper, rather than as a legal conclusion on what constitutes
scientific misconduct.  The discussion in McCaa makes it clear that the
hearing officer found that, while certain of the inaccuracies at issue
may have been merely negligent mistakes, others could only have been
made by Dr. McCaa with the knowledge that they would mislead the reader.
The hearing officer therefore specifically found that Dr. McCaa
presented the paper at issue with an "intent to deceive the reader."
McCaa at 35.

o       McCaa did not frame the issues in terms of whether scientific
misconduct had occurred (and could not have applied the PHS definition
of scientific misconduct, which was published in 1989, after McCaa was
issued).  The issues in McCaa were whether a cause for debarment had
been established and whether Dr. McCaa was "presently responsible" as a
recipient of federal funds.

o       The hearing officer found a cause for debarment in a pattern of
intentional deception by Dr. McCaa, which included fabricating data and
reporting as experimental procedures or results inaccurate information
based solely on expectations or suppositions.  Nothing in McCaa suggests
that negligent conduct alone is a sufficient basis for finding that a
researcher lacked integrity.

o       The expert testimony in McCaa related to the specific types of
inaccuracies at issue there, which were published in a scientific
journal article and may have misled readers in evaluating drugs used to
treat high blood pressure.  Thus, even if the testimony in McCaa had
established that negligent failure to prevent such inaccuracies would be
scientific misconduct, this would not necessarily mean that negligent
failure to prevent the inaccuracies in the grant applications at issue
here would also constitute scientific misconduct.

The unquestionable duty of a scientist to seek accuracy, which was
recognized generally and then applied to particular facts in McCaa, does
not necessarily extend to a duty to achieve perfection or freedom from
all inadvertent error. 5/  "Honest error" is expressly excluded from the
definition of scientific misconduct adopted by PHS, so something more
than inadvertence must be shown to establish scientific misconduct.

ORI gave virtually no notice in this case that it intended to prove
scientific misconduct on any basis other than intentional false
statements.  Despite passing references to what Dr. Sharma should have
known, ORI's Summary of its Case and its entire presentation at the
hearing were directed to its claim of intentional deception.  See, e.g.,
Tr. at 12-24 (ORI opening argument) and Tr. at 961-972 (ORI closing
argument).  Therefore, Dr. Sharma could not reasonably have been
expected to have presented any contrary evidence as to the standards
which ORI now claims that he negligently violated.

Nevertheless, ORI contended in its post-hearing brief that it could
establish scientific misconduct in this case by showing that Dr. Sharma
should have known that errors existed in the grant applications.  ORI
Post-Hearing Br. at 27.  ORI contended that it had established that a
principal investigator has a duty to "ensure the accuracy" of grant
applications, which Dr. Sharma failed to meet.

ORI's reliance, before us, for proof of negligence falling so far below
commonly accepted standards as to constitute scientific misconduct on
evidence of Dr. Sharma's general sloppiness in writing and proofreading
is particularly unpersuasive.  All the relevant facts were known to ORI
from the beginning of its investigation and yet no finding of scientific
misconduct was based on any of these items.  In fact, ORI specifically
found no scientific misconduct concerning other errors in the grant
applications and based its result on findings that:

 (1)  The statement on the à2GC probe retained in the revised
 grant application "was an editing error that resulted from
 inadequate application preparation and review procedures and
 does not constitute scientific misconduct."  ORI Final Report at
 54.

 (2)  Certain errors "resulted from poor supervisory practices
 and careless and sloppy proposal preparation.  Consequently, the
 statements . . . were unintentional errors and did not
 constitute scientific misconduct."  Id. at 55.

 (3)  A statement "resulted from uncritical integration of
 contributions from other team members, and that such conduct
 does not constitute scientific misconduct."  Id.

 (4)  "[O]ther misstatements in the Applications . . . were
 caused by careless yet honest errors - typographical, editing,
 and supervisory - and do not constitute scientific misconduct."
 Id.

Thus, ORI explicitly took the position in this case that sloppiness,
poor review procedures, and multiple errors, if unintentional, did not
constitute scientific misconduct.  Furthermore, ORI's scientific expert
testified that scientists vary in the degree of care with which they
proofread their grant applications.  Tr. at 614 (Dr. Douglass).  Dr.
Sharma testified that he did proofread these applications.  Tr. at 854.
The fact that his proofreading did not eliminate the errors that have
emerged from the close scrutiny to which these applications have been
subjected does not suffice to demonstrate that his conduct did not fall
within the range of practices accepted among scientists preparing grant
applications.  Not one of ORI's witnesses testified that unintentionally
submitting an application containing errors was considered scientific
misconduct.

ORI also argued that Dr. Sharma could have prevented the error by
providing the applications to other scientists for review before
submitting them, and that not doing so here was "tantamount to wilful
disregard" of their accuracy.  ORI Post-Hearing Br. at 28-29.  ORI
relied for its claim that such sharing was "a commonly accepted practice
within the scientific community" on the testimony of two scientists.
Id. at 28, n.72.  Dr. Douglass testified that he personally shares his
grants with various other researchers to get a critical review that may
allow him to improve the grant and get a better chance of funding.  Tr.
at 583-4.  Dr. Marshak testified that his practice was to have at least
two other investigators read his grant applications and that his
institution (Cold Spring Harbor) had a policy or at least a suggestion
for such review, but he did not know if it applied to senior
investigators.  Tr. at 199, 230-1.  Dr. Chalberg, who testified on
behalf of ORI, stated that he did not know at the time of any common
practice to have others review grant applications, although he thinks
that is now the rule.  Tr. at 118.  ORI itself found that CCF did not
have any clear policy requiring review of applications by other involved
scientists, and included a recommendation for adoption of such a policy
in its final report.  ORI Final Report at 56.  This evidence is not
sufficient to establish that the failure to have a grant application
reviewed by other scientists before submission would constitute a
serious deviation from a commonly accepted practice.  In any case, ORI
did not disprove Dr. Sharma's assertion that he did share the
applications with certain colleagues before submission. 6/  ORI merely
proved that he did not share final copies of them in advance of
submission with Dr. Chalberg, who now believes he could have caught the
errors.  See Tr. at 38, 77, 119 (Dr. Chalberg).

More rigor in proofreading and sharing the applications with peers might
well have avoided the present controversy. 7/  However, at the risk of
stating the obvious, we note that labeling conduct as "scientific
misconduct" must be viewed as a significant and consequential action:
notwithstanding the relatively minor corrective actions proposed in a
case such as this one, the opprobrium which naturally accompanies the
label has the potential to cause serious embarrassment and might even
destroy a scientist's career.  Thus, in the absence or law or
regulations establishing standards, we need to take considerable care in
assessing how substantially evidence shows a standard distinguishing
between merely sloppy work on the one hand, and, on the other, actions
which should produce the serious consequences that the brand "scientific
misconduct" entails.  It is against that background that we assess ORI's
arguments concerning whether such things as lack of care in proofing
documents rises to the status of "scientific misconduct."  We find that
no persuasive evidence was presented at the hearing or in ORI's
submissions to support the existence of a lesser standard applicable in
this case than one involving intentional falsification.  We therefore
limit the remainder of our analysis to a consideration of whether ORI
proved that Dr. Sharma intentionally made false statements in his grant
applications.

We turn next to analyzing the evidence which ORI presented to show
intentional deception in relation to each of the grant applications.  We
first discuss the à2GC grant application (Exhibit H-2), about which ORI
presented the bulk of its evidence, and then discuss the 180 kD protein
grant application (Exhibit H-3).

   Analysis of the Evidence on the Grant Applications

I.  The First Grant Application

The charges relating to Exhibit H-2 center on a parenthetical phrase in
one statement on page 21 of that grant application, which is highlighted
below:

 The uniqueness of the à2GC-receptor is further substantiated by
 our ongoing molecular cloning studies.  Utilizing a 39-oligomer
 probe (corresponding to the amino acid sequence of à2GC-receptor
 proteolytic fragment), which was synthesized using the rat codon
 bias data, three potential cDNA clones (1.4-, 2.8-, 4.5-kb) from
 2 million screened recombinants (þ ZAP rat adrenocortical
 carcinoma library) have been isolated and partially sequenced .
 . . .

The essential error in this sentence is that the probe actually used did
not correspond to à2GC sequence data but rather to sequence data of a
related receptor protein. 8/  The experiment described here was in fact
performed using a 39-oligomer probe which corresponded to the amino acid
sequence of an à2A receptor fragment.  Respondent Post-Hearing Br. at 2.
The à2A sequence data was obtained from human platelet sources and was
published by another researcher, R.J. Lefkowitz, in a paper in Science
in 1987.  Tr. at 880; Ex. R-13. 9/  We find that the sentence was
otherwise accurate.  The probe was synthesized using rat codon bias
data, as reported.  Tr. at 175-6.  It was 39-oligomer in size, as
reported.  Tr. at 152.  The use of the à2A probe resulted in recovery of
three potential clones after screening of the rat adrenocortical
carcinoma library, as reported.  See, e.g., Tr. at 880 (Dr. Duda).
Therefore, the error underlying the charges relating to Exhibit H-2 was
the substitution of à2GC for à2A in this clause. 10/

Dr. Sharma argued that this misstatement was the result of a single
typographical error in substituting à2GC for à2A in the underlined
portion of the sentence, thereby implying that he had amino acid
sequence data of à2GC which was used to construct the probe used in the
reported experiments rather than having relied on the published sequence
data for à2A.  Consequently, it is essential to examine whether this
subscript error was most likely to have occurred as a deliberate attempt
to mislead the grant reviewers, as ORI argued, or an unintentional or
clerical mistake, as Dr. Sharma contended.

We examine first what the record before us establishes as to how this
error occurred.  We conclude that, in the context of the full grant
application, this error was unlikely to and did not in fact mislead
reviewers, and that Dr. Sharma's conduct was inconsistent with
intentional deception.  On the other hand, we conclude that the
subscript substitution could most easily have been caused by a
typographical error.  We then turn to the arguments presented by ORI in
support of its contention that intentional deception nevertheless
occurred.

 o  The argument on which ORI relied most heavily wasthat Dr.
 Sharma offered inconsistent explanations of the error and that,
 while none of his statements constituted a plausible
 explanation, the inconsistency among them evidenced guilty
 intent.

 o       ORI also argued that Dr. Sharma's guilty intent was
 demonstrated by:

  (1)  his failure to cite the source of the probe
  immediately after the disputed statement,

  (2)  a pattern of misstatements that overstate his
  laboratory's achievements,

  (3)  his resistance to reporting the errors to NIH, and

  (4)  his lack of credibility.

We conclude that none of ORI's arguments persuasively support the
allegations of intentional deception in relation to Exhibit H-2.

 A. ORI's Theory of Intentional Deception Is Unlikely

ORI never fully developed a theory of its case which would explain how
Dr. Sharma allegedly sought to mislead the reviewers by his actions in
relation to Exhibit H-2.  However, ORI alleged generally that the motive
for intentional deception was that the reviewers would perceive the
research as more advanced if they believed that Dr. Sharma had already
obtained amino acid sequence data for a segment of the receptor protein
he was studying, i.e., à2GC.  ORI Post-Hearing Br. at 10-11 and record
cites therein.  There is no serious dispute that the possession of such
data on à2GC would have been an attractive feature in the grant
application, because a probe based on à2GC sequence instead of à2A was
at least somewhat more likely to retrieve clones of à2GC receptor
protein 11/ and because the data could be used later to help verify the
identity of the clones obtained.  See, e.g., Respondent's Post-Hearing
Br. at 2-3.

However, credible evidence in the record showed that an achievement such
as obtaining amino acid sequence data on the specific receptor protein
(i.e., à2GC) which was the subject of the research would have been
highlighted in the grant application.  Thus, one scientific expert with
experience as a grant reviewer, Dr. Gelmann, testified that --

 it would be a major accomplishment and in a grant like this,
 somewhere there would be several pages of detailed description
 about isolating that fragment, purifying it and deriving the
 sequence.  And that's a major piece of work that could very well
 take more than a year.  So when I read this the first time, it's
 obvious that he doesn't have this and that's got to be a
 mistake.

Tr. at 931. 12/  He also stated that the description in the preceding
paragraph of the grant application about Dr. Sharma's "analysis of the
putative Alpha 2GC protein" makes clear that the "analysis is far too
preliminary to get him [Dr. Sharma] to a proteolytic fragment that he
could sequence."  Id. 13/  So the record shows that, without giving the
data and explaining how he obtained it, Dr. Sharma was not likely to
succeed in misleading the reviewers by the single reference to à2GC
data, especially since other material in the grant implied that such
data was not available. 14/

Nevertheless, ORI argued that the reviewers were, in fact, misled.  ORI
relied on the following quote from the critique prepared by the study
section that reviewed the application:

 The applicant has evidence that the receptor under investigation
 is structurally and functionally distinct from two subtypes of
 alpha-2 adrenergic receptors identified in humans . . . .
 Progress in the cloning of the gene for the alpha-2 adrenergic
 receptor has been less impressive and the success of a different
 laboratory in cloning and sequencing the closely related alpha-2
 adrenergic receptor from human platelets has obviously
 diminished the appeal of the cloning studies.

ORI Post-hearing Br. at 12; Ex. R-9, at 2.  Neither the quoted language
nor its context will bear ORI's construction.

Nowhere in the report does the study section point to an achievement of
obtaining sequence data on à2GC, even though the report recommends
approval of the grant application "with enthusiasm" and would thus be
expected to point to an achievement that would help justify its
conclusion.  Rather, the study section acknowledges a weakness of the
application in that Lefkowitz' à2A research was more advanced than Dr.
Sharma's à2GC research, since à2A had already been cloned and sequenced.
If the study section thought that Dr. Sharma already had at least
partial sequence data on his à2GC receptor, they would have said that
the other laboratory had succeeded in obtaining full amino acid sequence
data on its receptor by way of contrast.  As it stands, the sentence
implies that the study section understood that Dr. Sharma's laboratory
had not had success in sequencing à2GC.  In the portion which was
omitted in ORI's quote, the critique refers to the "significant strength
of the proposal" in Dr. Sharma's "demonstrated success . . . in
purifying and biochemically characterizing membrane receptors and
enzymes" in these cells.  Ex. R-9, at 2.  The study section does not
cite as a strength any success in obtaining à2GC sequence data.  The
study section critique thus supports the conclusion that the single
reference to à2GC sequence data in the parenthetical of the sentence at
issue was not enough to mislead the reviewers into believing that Dr.
Sharma had actually obtained such data.

ORI also suggested that no evidence supported the statement that the
à2GC receptor was "structurally" distinct from other related receptors
unless the study section believed that Dr. Sharma had amino acid
sequence data for à2GC.  ORI Post-Hearing Br. at 12.  ORI cited nothing
in the record to support this assertion.  While it is obvious that
possession of the complete amino acid sequence would be primary proof of
the structure of the protein, ORI did not prove that less conclusive
data (such as proteolytic cleavage results, peptide mapping, and
chemical analyses) could not be relied on to deduce some information
about the structure of the à2GC protein and some evidence of its
distinctness from related proteins (such as à2A). 15/  Thus, ORI did not
prove that the critique's reference to Dr. Sharma's having evidence of
structural distinctness demonstrated that the study section believed him
to have amino acid sequence data.

We conclude that, although a scientist focussing only on the erroneous
statement might well assume that à2GC amino acid sequence data was
available, a reviewer who looked at the entire application could not be
expected to accept that reference at face value. 16/  Thus, a single
reference in a parenthetical to making a probe using such data risked
being viewed as an error or at least a non sequitur, or of simply being
overlooked by the study section reviewers.  ORI failed to provide any
proof that the single reference to à2GC could plausibly have been an
attempt at deliberate deception.

If the inclusion of the erroneous reference to à2GC was intentional at
all, it must have been an extremely subtle deception which was
intentionally placed in a relatively obscure place.  Yet Dr. Sharma's
conduct was inconsistent with that of someone who was attempting to
conceal a false statement which he had knowingly inserted in an
inconspicuous manner.  For example, leaving the erroneous statement in
the revised grant application (Exhibit H-35) (where ORI concedes that it
was of no value to the methodology proposed), 17/ even though he deleted
many other statements in editing the revision, is patently inconsistent
with intentional deception.  Cf. ORI Final Report at 33.  If he had
consciously planted the reference in the original grant application,
surely he would wish to remove it immediately in his revision once it
had lost any utility and become, if anything, counter-productive.  Also,
his confused and garbled responses to the various investigators
(discussed in detail in relation to ORI's claims of inconsistent
statements) are not consistent with the strategic planning required by
the intentional deception scenario.  If he had been so guileful as to
intentionally plant an inconspicuous reference in order to be able to
deny it if its falsity were noticed, surely he would have prepared a pat
excuse.  Instead, the impression throughout is of a man baffled and
upset by an error he cannot explain to even his own satisfaction. 18/
After observing Dr. Sharma's demeanor, we do not find it credible that
he planned so subtle an intentional manipulation.  We find it far more
plausible that he made an honest, if careless, error.

We therefore conclude that, after looking at the context of the
statement in the application, the reviewers' responses, and Dr. Sharma's
conduct, intentional deception is not supported by the record as a
likely explanation of the subscript error.

 B. Typographical Error Is More Likely

A preconception existed both at CCF and at ORI that the transposition of
à2A and à2GC could not be explained by simple mistyping because they
required different keystrokes located far apart on a standard keyboard.
Tr. at 119 (Dr. Chalberg), 431-2 (Dr. Ferrario), and 702 (Dr. Bivens).
However, at the hearing, Dr. Sharma's wife explained that she had
created macros for the laboratory computer program in which à2A and à2GC
differed by a single keystroke.  In that case, a simple typographical
error is the most plausible possibility.

ORI responded that Dr. Sharma never told anyone about the macros before,
and that he always said he typed the entire application.  ORI
Post-Hearing Br. at 19-20.  The record cited by ORI directly contradicts
this characterization.  Dr. Sharma told ORI that his wife helped him
whenever he needed help, that it was very difficult for him to use the
computer to create the necessary subscripts and Greek letters, and that
she would come in and either show him how to do it or go ahead and do it
herself, and that she did formatting of portions of the text which he
marked for her attention.  Ex. H-31, at 220-24 (Interview of Dr.
Sharma). 19/  Mrs. Sharma's role in helping with the typing of grant
applications on the computer word processor was also corroborated by a
letter from Dr. Chalberg to OSIR.  Ex. H-17, at 2.

Dr. Chalberg admitted making an identical mistake in the same grant
application by substituting à2GC for à2A in a sentence reporting on
Lefkowitz's research and asserted that his was a typographical error.
Tr. at 160.  Obviously, he had no motivation to misrepresent the work of
Lefkowitz as dealing with à2GC rather than à2A, and ORI did not suggest
that this admittedly false statement constituted scientific misconduct.
His lack of intent was apparently assumed in the absence of any evident
benefit from the misstatement.  While the lack of benefit or motive is a
reasonable basis for finding the absence of intent, the occurrence of an
identical error in another section of the application for which a
benefit or motive can be assigned is not alone a sufficient basis for
finding intent, as ORI suggested in relying on the fact that Dr.
Sharma's error occurred in the methodology section rather than merely
the background.  ORI Post-Hearing Br. at 18.  At the least, the
admission that Dr. Chalberg made such a typographical error corroborates
the testimony of Dr. and Mrs. Sharma that such an error was not
difficult to make under the system of macros set up in the laboratory's
computers.  See Tr. at 801-6 (Mrs. Sharma testified that the macros were
in a glossary to which others working on the research plans would have
had access.  Tr. at 802.).  ORI itself recognized that the occurrence of
this similar error in a context that was "clearly typographical" lent
some support to Dr. Sharma's claim that the substitution on page 21 was
similarly unintentional.  Compare ORI Final Report at 26 with ORI
Post-Hearing Br. at 19. 20/

In fact, there have been no shortage of admissions and examples showing
the ease with which the two adrenergic receptors can be accidentally
transposed.  See, e.g., Tr. at 618-23.  For example, ORI itself twice
asserted that "changing 2A to 2GC would not cure the false statement" at
page 21 of Exhibit H-2, when that statement already reads 2GC and ORI
obviously meant changing 2GC to 2A would not suffice to correct it.  ORI
Post-Hearing Br. at Table of Contents and 14.

ORI acknowledged in its Final Report that the possibility of a
typographical error was enhanced by the frequency with which references
to à2GC and à2A recurred in Exhibit H-2, which ORI estimated as a total
of about 128 references.  ORI Final Report at 26.  ORI also acknowledged
the Dr. Sharma's lack of expertise in computer word processing at the
time could have played a role.  Id.

We conclude that the subscript substitution most likely occurred through
a typographical error.  We next discuss the several arguments offered by
ORI in an effort to demonstrate that the misstatements nevertheless were
not the result of honest error.

 C. ORI's Arguments On Dr. Sharma's Intent Are Not Supported on
 the Record

  1. The Alleged Inconsistency of Dr. Sharma's
  Explanations

   a.  General considerations

Dr. Sharma admitted that he did not have à2GC amino acid sequence data
from which to make a probe, and that he did not make a probe using such
data, so the statement referring to such data in Exhibit H-2 is
unquestionably erroneous.  See, e.g., Respondent Post-Hearing Br. at 1.
Beyond his admission of error, Dr. Sharma made a number of statements to
Dr. Chalberg, Dr. Ferrario, the CCF inquiry committees, and ORI
interviewers, which ORI interpreted as explanations for how this error
occurred.  Although ORI did not accept as accurate any of the supposed
explanations, it considered them to be inconsistent with each other and
to undermine Dr. Sharma's credibility, and therefore to be probative of
intentional deception.

We will look at the specifics of these explanations below.  Generally,
we find that the various statements, viewed as a whole, do not support
ORI's contention that Dr. Sharma intentionally inserted the reference to
à2GC in order to mislead reviewers as to the status of his research.  In
many cases, the questions to which Dr. Sharma was responding differed,
or were unclear and seemed to call for speculation, or did not relate to
either of the grant applications at issue here. 21/  In some cases, a
statement which ORI highlighted as inconsistent had quite a different
meaning when read in context.  At times, Dr. Sharma's statements were
confusing and susceptible of more than one interpretation.  However,
none of the statements, alone or read together, was consistent with
intentional deception.

In weighing Dr. Sharma's prior statements, we considered several
factors.  First, in observing Dr. Sharma's demeanor at the hearing, we
found him to be credible in responding to the instances of apparently
inconsistent prior statements.  Second, it was evident, both from
observing his responses at the hearing and from reviewing the
transcripts, tapes, and reports of prior interviews with him, that Dr.
Sharma does not communicate with facility in English (which is
apparently not his first language), despite his general familiarity with
the language and his proficiency in some technical areas.  The Panel
found that he did not appear to understand fully some questions to which
he attempted to respond and that he did not always use language with
precision.  Therefore, we find it believable that some of the prior
communications may have resulted from his misunderstandings of the
questions or misunderstandings of his intended responses.

Furthermore, since the burden of proof in this matter rests with ORI,
the significance of any of Dr. Sharma's ambiguous statements had to be
measured in relation to the purposes for which ORI was offering them.
ORI did not argue that any of the statements amounted to an admission of
what ORI contended actually occurred.  At most, ORI attempted to show
intent in a very indirect way, by arguing that Dr. Sharma offered so
many different explanations that the true explanation must be
intentional deception.  For this point, ORI must at least prove that the
explanations offered did directly contradict each other and could not be
understood in a consistent manner.  Merely showing that repeated
questioning over a number of years elicited various muddled answers does
not suffice to support ORI's claim of intentional deception.

Overall, we find that Dr. Sharma has asserted throughout that he did not
intentionally insert the incorrect statement in Exhibit H-2, and has
asserted that therefore it must have been substituted by an error either
through his typing or through his failure to catch it in editing.  The
statements which might be considered inconsistent with this position
were references to anticipating receipt of à2GC sequence data.  ORI
considered these statements as admissions that Dr. Sharma intentionally
wrote in à2GC in the disputed sentence in the hope that he would get
à2GC sequence data in time.  We find it unreasonable to read these
statements as intended by Dr. Sharma to mean that he had inserted à2GC
because he hoped to get the data in time or to remove it if necessary
(so-called "anticipatory writing"). 22/  Anticipatory writing makes
little sense, in light of the structure of the sentence in which the
reference to à2GC occurs on page 21 of Exhibit H-2.  Even if the
sequence data on à2GC had arrived shortly before or after the grant was
submitted, the sentence would be no more accurate than it is now.  The
sentence does not simply report the receipt of sequence data; it reports
the use of that data to construct a probe, the results of using that
probe to screen a specific library, and the particular weights and
restriction patterns of the resulting cDNA clones.  If the data were
received, new experiments would have to be conducted and the entire
sentence replaced with a new sentence reporting the presumably different
results of those experiments.  Dr. Sharma would have been more aware of
this than anyone else and would also have known that this would quickly
become obvious to any scientist who looked closely at the sentence.  He
would have had to be quite unthinking to have proffered such an
explanation.  In fact, ORI and others reviewing the case recognized that
Dr. Sharma could not have inserted à2GC as anticipatory writing.  See,
e.g., ORI Final Report at 19, ORI Post-Hearing Br. at 22-24; Ex. H-7, at
4.  Nevertheless, ORI viewed such "admissions," however implausible, as
evidence of some deliberate intent to deceive.

Upon close review of those statements, we find that Dr. Sharma was
making three quite different points: 23/

 --  First, that, rather than simply brush the error off as a
 "typo," he wanted to try to understand how he could have made
 such a mistake or how the investigators could think he acted
 intentionally.  Thus, he speculated that his constant concern
 about trying to obtain the sequence data on à2GC might have
 caused him unconsciously either to mistype the subscript or to
 fail to notice it in proofreading, which he called
 "psychobiology."  See Ex. H-31, at 272-3, 292, 320; Tr. at
 828-9.  In other instances, he speculated that the investigators
 might believe he was anticipating data in order to justify the
 sentence when that was not possible.  See Ex. H-31, at 298.

 -- Second, that he answered various questions about whether he
 anticipated, expected or needed the data in the affirmative,
 because he did want and hope to obtain à2GC sequence data
 throughout the relevant time period.  Tr. at 820, 834.  However,
 he needed the data for purposes of verifying the identity of the
 clones (which he had already obtained using the à2A probe), not
 for purposes of constructing a replacement for the probe
 referred to in the relevant statement.  Tr. at 821.  None of the
 colloquies in prior interviews which ORI treated as inconsistent
 with this position were clear in asking Dr. Sharma whether he
 anticipated the data for purposes of constructing the probe.

 -- Third, that he intended to remove the sentence from the
 revised grant application (Exhibit H-35) altogether, because he
 had changed course so that it was irrelevant to the revised
 methodology in that grant, and had missed it.

ORI was aware, as noted above, that a theory of anticipatory writing
would not make sense, and considered these possible interpretations of
Dr. Sharma's statements but rejected them as not credible.  ORI Final
Report at 30.  We reach the opposite conclusion for the reasons
discussed more fully below.  Finally, as we discuss in detail below, we
discount some of the reports about Dr. Sharma's statements because of
the atmosphere of conflict and mistrust which existed between Dr. Sharma
and the persons reporting those statements, in particular Dr. Chalberg
(the original accuser who worked in Dr. Sharma's laboratory) and Dr.
Ferrario (Dr. Sharma's former department head).

  b. Specific allegedly inconsistent statements 24/

   i. Statements on "anticipatory writing"
   duringinvestigation

The Healy Committee report of its interview with Dr. Sharma described
him as saying that at "the time he was writing the grant, he was
anticipating that at any moment he would have had the sequence to
fashion the probe.  When the sequence data did not come, he changed
course and decided to make the probe based on" à2A.  Ex. H-5, at 3.  ORI
referenced this report as evidence that Dr. Sharma had told the Healy
Committee that he made "an editorial error" in not removing the
reference to à2GC from the first grant application when he did not get
the data.  ORI Post-Hearing Br. at 34-35.  However, ORI's brief ignored
the next sentence in the Healy report, which states that Dr. Sharma
"pointed out that in the actual Progress Report section of the grant, he
explained exactly what he had done, which was: `isolated a cDNA clone
from rat brain using human platelet alpha2-adrenergic receptor genomic
clone as a probe . . ."  Ex. H-5, at 3.  Since this sentence refers to a
cDNA clone rather than a 39-oligomer probe, and to rat brain cells
rather than rat carcinoma cells, this sentence obviously refers to the
research methodology of the later revised application, and not to the
first application at issue here.  Dr. Sharma is reported to have said
further that "he should have taken the Page 21 statement out because he
actually used a different approach and that he overlooked correcting
that error and it was clearly a careless mistake and was not
intentional."  Id.  Evidently, the Healy Committee was not clear about
the change in methods from the first (using an à2A 39-oligomer probe) to
the revised version (using a genomic probe) and understood incorrectly
that the change to which Dr. Sharma referred was from a plan to make an
à2GC probe to the actual use of an à2A probe. 25/

ORI argued that it is immaterial that Dr. Sharma was responding (both
before the Healy Committee and at other points in the investigation) to
the reference to à2GC in the revised application (Exhibit H-35) rather
than in the unfunded application at issue here (Exhibit H-2).  ORI
contended that, if Dr. Sharma included the reference in Exhibit H-35
"because he hoped to get it [the sequence data] before it [Exhibit H-35]
was reviewed, then he must have also purposely included it in
Application 1 which he submitted months before."  ORI Post-Hearing Br.
at 33, n.97, and 20, n.52.  This argument entirely misses the point.
When Dr. Sharma stated (to the Healy Committee, for example) that he had
"changed course" and "overlooked correcting the statement" when he
revised the application, it is clear that he is saying that the entire
statement, not just the reference to à2GC, did not belong in that grant
application, "because he actually used a different approach," i.e.,
using a cDNA clone from rat brain cells.  See Ex. H-5, at 3.  He thus
failed to "catch" this statement and remove it, as he had removed all
the other disputed statements in Exhibit H-2 from Exhibit H-35.  For Dr.
Sharma to state that he meant to take the entire statement out in
revising the grant to change its methodology does not imply that he knew
that the statement contained the misstatement at issue, and far less
that he intentionally included the subscript error when he first wrote
Exhibit H-2.

Perhaps the most troubling of the statements relied on by ORI was
recorded during the interview of Dr. Sharma by the Hamilton Committee.
26/  Dr. Sharma stated:

 Now the question comes, why in my own mind I did not -- this is
 incidentally a 48-page document. 27/  That is the only omission
 which is in 48 document [sic] that GC has not been converted to
 alpha two-A, and still a question there was in people's mind.
 Maybe Ramesh is trying to be sneaky, and is trying to pull a
 fast one because we cannot prove, we cannot get him, and maybe
 he still had a bad intent. . . . Originally, when I wrote really
 the GC -- and I am again hurting myself by saying it -- my
 intention was at that time, when I had put it in, that I might
 get a sequence of GC because if I don't get a sequence of GC,
 then I will change it to alpha 2A.  That takes very little part
 in the computer.  Now am I making that statement correct?

Ex. H-29, at 68-69. 28/  The context of the statement was again a
discussion of the revised grant relating to rat brain research in which
"this statement got left" in the process of revising it to meet the
criticisms of the study section.  Id. at 67, 71-3.

Dr. Sharma explained when confronted with this statement by ORI that he
was speculating how people might think he behaved in a "sneaky" way,
despite the fact that only the one error occurred in the sentence at
issue and only that one sentence was erroneously left in the revised
grant (Ex. H-35).  Ex. H-31, at 297-98.  Similarly, at the hearing, he
described this statement as "hypothesizing the situation," since Dr.
Hamilton had told him to speculate.  Tr. at 828.  ORI denied that Dr.
Sharma was told that "the norm of the day was speculation."  ORI
Post-Hearing Br. at 36.  However, in fact, the following colloquy
occurred earlier in another context in the same Hamilton Committee
interview:

 MS. HARTER:  Can you think of any reason -- what possibly
 prompted you to maybe --

 DR. HAMILTON:  Speculate?

 DR. SHARMA:  That is a very good question . . . .

Ex. H-29, at 19.  It was thus not unreasonable for Dr. Sharma to believe
that he was expected to speculate about how errors might have occurred.

Dr. Hamilton himself was not sure what Dr. Sharma meant.  When Dr.
Hamilton was asked at the hearing by ORI counsel whether there was "any
doubt" in his mind that Dr. Sharma told his committee that "the a2GC was
put in the February grant application, because he was anticipating
receiving that data," he responded that he "would not put it that way."
Tr. at 446.  Instead, he said he had "no doubt that he [Dr. Sharma] was
anticipating that data and that that had some relationship to his having
put the term `a2GC' in there."  Id.  He was certain that Dr. Sharma did
say he "anticipated having that sequence available," but concluded:  "I
can't really testify as to his intent."  Tr. at 447. 29/  Dr. Hamilton's
first-hand impression about these statements is not inconsistent with
the position that Dr. Sharma was indeed anticipating sequence data, and
that this anticipation played some role, perhaps unconscious, in his
error, but that Dr. Sharma did not necessarily assert that he
consciously wrote à2GC in the hope of getting the data in time to submit
the grant.

Dr. Hamilton also testified that Dr. Sharma pointed out first to his
Committee that the "only error there was the inclusion of the letters
`GC.'"  Tr. at 355.  He stated that Dr. Sharma said that, if he had
referenced à2A, "then the statement would not have been in error.  And
he took us through the entire grant, pointing out that, in fact, that
was true."  Id.

In other parts of the Hamilton Committee interview, Dr. Sharma made many
statements related to having made a typographical or computer error in
substituting à2GC.  See, e.g., Ex. H-31, at 295 ("I did not know I put
gc in there.").  Thus, Dr. Sharma said:

 [A]t that time the secretarial help available to us was not very
 much -- and I am not a computer wizard . . .  When I was typing
 these grants, the only misstatement, if you want to say -- and
 that's not a deliberate statement -- is on page 21. . . .  I
 don't want to buck the issue just by saying it's a typo and cut
 it over.  No, that's not the case I'll address that issue.  It
 should have been alpha two A instead of alpha 2GC.

Ex. H-29, at 52-3.  It is reasonable to interpret the point of this
statement to be that the error was caused by typing or computer
mistakes, but that Dr. Sharma does not wish to appear to be making too
facile an excuse.   Thus, in his comments to the Hamilton Committee,
either Dr. Sharma was indeed foolishly making opposite statements
simultaneously to the same people or he was being misunderstood.

Further, Dr. Sharma clearly understood in talking to the Hamilton
Committee that anticipatory writing was not a possible explanation for
the error in the original or revised grant.  Dr. Morton, a committee
member, pointed out that changing the subscript to 2A in Exhibit H-2
would not make the statements about the results of the probe correct.
Dr. Sharma responded twice that that was right, and that "in the
original grant which I have made . . . it is very easy once you have the
sequences to confirm the identity of the clone.  And that thing was
always at the back of my mind that ultimately we have to confirm that
alpha 2GC is really alpha 2GC or not, and we need the sequences for
those.  But I thought I could change it."  Tr. at 74.

Dr. Sharma's main point in this statement is reasonably understood to be
that he would be able to use the sequence data to justify his claims
that the potential clones (which he had already obtained with the à2A
probe) were of à2GC.  These claims were based largely on his biochemical
and pharmacological studies about adrenergic receptors in the rat cells
he was researching but on relatively little structural data.  Tr. at
952-3.  The identity of the clones as à2GC could be substantiated by
comparison to à2GC sequence data once it became available, as well as by
expression of the protein.

In summary, we find that Dr. Sharma's statements to the Hamilton
Committee were unclear in places.  However, Dr. Hamilton was not himself
certain whether Dr. Sharma was telling him that he wrote the reference
to à2GC because he anticipated getting the sequence data.  Overall and
in context, we find that the statements do not significantly undermine
Dr. Sharma's credibility or support ORI's contention that Dr. Sharma
deliberately inserted the reference to à2GC sequence data on page 21 of
Exhibit H-2.

The Fishleder Committee report did not refer to any anticipatory
explanations and found that the misstatement on page 21 of Exhibit H-2
"arose because of a single typographical error."  Ex. H-8, at 3.  Dr.
Sharma reiterated the assertion that the misstatement was the result of
a typographical error numerous times.  For example, in correspondence
with OSI, Dr. Sharma repeated that the errors in the original grant
application were "the result of a single computer-typographical error at
a single place in a subscript on page 21:  there should have been à2A
instead of à2GC. . . .  The above mentioned 39-mer probe corresponding
to the à2A-receptor proteolytic fragment was synthesized by Dr. Michael
Cashnell . . . ."  Ex. H-20, at 2; Ex. H-21, at 2-3.  In regard to the
revised application, he stated in the same letters that the erroneous
statement "was directly reproduced from the computer and therefore
escaped editing."  Ex. H-20, at 2; Ex. H-21, at 3; see also Ex. R-2, at
1-3 (Comments of Dr. Sharma to OSIR); Ex. H-10, Attachment 1, at 2-3.

   ii. Statements reported by Dr. Chalberg

Dr. Chalberg (the researcher in Dr. Sharma's laboratory who initially
raised these allegations) testified that he looked at the revised grant
application (Exhibit H-35), 30/ and noted the reference to à2GC sequence
data as the source of a probe.  He stated that he pointed this out to
Dr. Sharma and was told that they did have the sequence data.  Tr. at
32-36.  Dr. Chalberg stated that he pointed out that even if the data
had been received, he would have known if a probe had been constructed
based on that data and no such probe was made.  Tr. at 37.  According to
Dr. Chalberg, Dr. Sharma did not reply to that point, but later showed
him a version of the grant without the disputed statement.  Tr. at 37-8.
Dr. Chalberg also testified that he was asked by Dr. Ferrario to look at
Dr. Sharma's grants in order to identify his own contributions as part
of his performance evaluation, and discovered in the process that the
erroneous statement about the source of the probe appeared in the
official copy of the revised application (Exhibit H-35).  Tr. at 38-51.
He was instructed to ask Dr. Sharma about the matter again, and
testified that Dr. Sharma told him that he [Dr. Chalberg] "still did not
have the correct version of the grant, that the version he [Dr. Sharma]
sent did not have that statement in it."  Tr. at 71-72.

The record demonstrates that longstanding conflict existed between Dr.
Sharma, Dr. Chalberg, and Dr. Chalberg's fiancee, a technician in Dr.
Sharma's laboratory.  See, e.g., Tr. at 101-3, 160-1, 438-9; Ex. H-29,
at 82-84, 94-102.  Dr. Chalberg was already trying to leave the
laboratory and there is some evidence that he and his fiancee were
exploring working under Dr. Ferrario instead of Dr. Sharma.  Tr. at 148;
Ex. H-5, at 4-5 (Healy Committee Report).  ORI called as a witness Mr.
Michael Murray, a non-scientist administrator from CCF who was asked to
sit in the meeting between Dr. Ferrario and Dr. Sharma.  Mr. Murray
reported that Dr. Chalberg's central concern was that his name was on
the grant applications.  ORI based its claim that Dr. Chalberg acted
with disinterested motives on this assertion.  Tr. at 271; ORI
Post-Hearing Br. at 39, n.124.  However, Dr. Chalberg does not appear
among the personnel listed in Exhibits H-2 or H-35, relating to
adrenergic receptors, and Dr. Sharma testified that Dr. Chalberg's work
was in the area of the molecular biology of 180 kD protein.  Tr. at 870.
31/  ORI presented no rebuttal testimony.  Since the initial charges Dr.
Chalberg brought to Dr. Ferrario related to the Exhibit H-35
application, the explanation that Dr. Chalberg happened to discover the
misstatement while reviewing the applications for his own performance
evaluation, and acted to protect his own reputation, is questionable and
calls his credibility into question.

However, even assuming that Dr. Chalberg was reporting the conversations
to the best of his recollection, it is not impossible that communication
between Dr. Sharma and Dr. Chalberg at that point may have generated
more heat than light.  It is also possible that emotion and language
difficulty contributed to lack of complete understanding.  In any case,
questions to Dr. Sharma from Dr. Chalberg regarding statements about an
à2GC 39-oligomer probe in the revised application would likely have
resulted in confusion.  Dr. Sharma was unlikely to know what Dr.
Chalberg was talking about, if the statement was retained in that
revision by unintentional error, as ORI found.

Dr. Chalberg's account is contradicted by Dr. Sharma's description of
the conversation, which he testified occurred in the presence of Mrs.
Sharma (as she corroborated).  See Tr. at 817-8, 807-8.  Dr. Sharma
testified that Dr. Chalberg stated generally that he had a problem with
"a statement about alpha 2GC" in one of Dr. Sharma's grant application,
but that Dr. Chalberg did not identify a specific statement in a
particular application.  Tr. at 817-18.  Dr. Sharma denied telling Dr.
Chalberg that he had received à2GC data or that the problematic
statement did not appear in another version of the grant application.
Notably, Dr. Chalberg is the only witness to have testified that Dr.
Sharma ever claimed to have received the à2GC amino acid sequence data,
while all the others, including Dr. Ferrario, agree that Dr. Sharma
admitted from the beginning that he did not have that data.  See, e.g.,
Exs. H-4, at 2, and H-5, at 2.

In light of their relationship at the time, Dr. Sharma may well have
reacted defensively to an accusation by Dr. Chalberg that something
about the 39-oligomer probe for à2GC was wrong in the revised
application which was funded.  However, it is not credible that Dr.
Sharma would have asserted to a researcher in his own laboratory that
they had recently received sequence data when (1) the grant application
was submitted about six months earlier (Exhibit H-35 is dated October
24, 1989) so recent receipt of the sequence data would be irrelevant;
(2) the grant about which Dr. Chalberg was speaking did not depend on
using an oligomer probe corresponding to amino acid sequence from either
the à2A or à2GC receptors; (3) Dr. Chalberg would have ready access to
information in the laboratory to challenge such an assertion; and (4) it
would have been obvious to both of them, if they looked at the specific
sentence, that receipt of à2GC sequence data would not in itself make
the sentence true since all the results reported were those obtained
with a à2A probe.  Dr. Chalberg himself pointed out that he would have
been well aware of it if the data had been obtained, at least if any
probe were going to be made using it.  We therefore conclude that Dr.
Chalberg's reports of Dr. Sharma's explanations are not credible or
accurate.

   iii. Statements reported by Dr. Ferrario

Dr. Chalberg testified that he took his concerns to Dr. Ferrario as head
of the department on April 20, 1990.  Tr. at 68-71.  Dr. Ferrario
testified that Dr. Sharma came to his office later that day and was told
about "the situation."  Tr. at 402.  Dr. Ferrario testified that he did
not have expertise in molecular biology and relied on Dr. Chalberg to
understand the significance of the charges.  Tr. at 406, 437.

Dr. Ferrario testified that, at the April 20th meeting with Dr.
Bannerjee and Mr. Murray present, Dr. Sharma told him that he wrote à2GC
in the application they were discussing, i.e., Exhibit H-35 (see Tr. at
422-5), because he anticipated getting the necessary data before the
grant was reviewed.  Tr. at 402-3, 425-6. 32/  Dr. Ferrario stated that
his "recollection is that he [Dr. Sharma] initially stated that he had
prematurely written in the grant data that was not available at that
time, but that he expected to have it before the grant was actually
reviewed" and that Dr. Sharma stated in regard to the receptor sequence
data that he did not have it "at that time, but that he expected to have
it."  Tr. at 402-3.  He insisted that Dr. Sharma did not characterize
the errors in the grant application as typographical at the April 20th
meeting, although that explanation "developed during the following
days."  Tr. at 425-33.  He explained, however, that "typographical
error" meant to him only transposing letters or misspelling words, while
"errors" would be "saying that I will do A when I really tried to do B.
But it takes me a few more words to say A than to say B."  Tr. at 431-2.
Therefore, he explained that, although he had earlier stated that Dr.
Sharma had said at the meeting "that there were mistakes in retyping the
grant and that he had not caught the error," that statement by Dr.
Sharma would not "imply a typographical error."  Tr. at 430-1.

Dr. Ferrario's interpretation is based on the preconception, noted
above, that was shared by Dr. Lyle Bivens, who directed the ORI
investigation, that the mistake here could not have been a classic
typographical error because the characters "2A" and "2GC" require quite
different keystrokes in typing.  Tr. at 702.  This assumption was
undercut by the uncontradicted testimony of Mrs. Sharma, discussed
above, that she set up a macro system for typing à2A and à2GC that
differed by only one neighboring keystroke.  Tr. at 801-6.  We thus
conclude that Dr. Ferrario's understanding that Dr. Sharma was not
talking about a typographical error on April 20th may have resulted from
this faulty assumption.  See Tr. at 433. 33/

Also, in weighing the context of Dr. Ferrario's conversations with Dr.
Sharma, we note that the record reflects that a high level of conflict
existed between the two.  See, e.g., Ex. H-5 (report of CCF preliminary
inquiry noting "contentious relationship . . . over the issue of Dr.
Ferrario's hiring Dr. Sharma's molecular biology technician, Janet
Rhine, and possible hiring of his Project Scientists Dr. Rama Jaiswal
and Dr. Chalberg"); ORI Final Report at 44; Ex. H-31, at 246, 249,
340-1. 34/  Dr. Ferrario considered Dr. Sharma "more difficult to talk
to or to work with than with other colleagues in the department."  Tr.
at 441.  The CCF investigations dealt with Dr. Sharma's complaints
against his treatment by and contract disputes with Dr. Ferrario, as
well as with the allegations of scientific misconduct against Dr.
Sharma.  See, e.g., Ex. H-6; ORI Final Report at 44-45. 35/  (At times,
Dr. Sharma's statements at earlier stages of this case seem to have
reflected his focus on his own agenda, rather than on the issues now
before us.  See, e.g., Ex. H-31, at 235-7, 252, 340-1.)

It is, of course, impossible to know exactly what was said in the
meetings which were not recorded.  We do not imply that Dr. Sharma's
complaints necessarily had merit or that Dr. Ferrario was less than
honest in reporting his memory of the meeting either in his memorandum
at the time (Exhibit H-4) or in his testimony at the hearing.  However,
the level of conflict, the admitted difficulty he had in talking to Dr.
Sharma, and Dr. Ferrario's lack of familiarity with the science involved
all make communication less reliably accurate and misunderstanding more
likely.

Ultimately, we conclude simply that the reports from Dr. Ferrario and
Dr. Chalberg of conversations with Dr. Sharma do not credibly support
ORI's claim of intentional deception.

  iv. Statements in letter from Dr. Sharma's counsel

Counsel for Dr. Sharma at an earlier stage of the investigation wrote a
letter which added to the confusion by denying that Dr. Sharma had said
he made a typographical error.  Ex. H-24.  The letter does so in less
than crystal clear terms by saying that "an editing error and
anticipatory writing are one and the same.  An editing error occurs if
one anticipates receiving the results of sequencing data and upon
learning that those results are not available inadvertently fails to
modify the text accordingly."  Ex. H-24, at 4.  On the other hand, the
letter distinguishes between saying that à2GC was "mistyped" and saying
that it was a "typographical error."  Id.  The letter argues that Dr.
Sharma's statements to the Hamilton Committee can be read as consistent
with an inadvertent failure to edit references after failing to obtain
the data, and not as assertions of typographical error.  Id. at 4-5.

When questioned about this letter, Dr. Sharma testified at the hearing
that:  "I don't review all legal things what people want. . . .  They do
not register.  These are not the normal days for me. . . .  I probably
reviewed it, but it doesn't register in my mind what is happening on
legal term."  Tr. at 855.  Shown the letter, Dr. Sharma commented
"[w]hat I am reading, I still don't understand what the question is and
what I am reading."  Tr. at 857.  Finally, he concluded with the
following exchange:

 A  I don't understand the difference between the editing error
 and the typing error.  So I cannot comment on this.  To me I had
 very little differentiation when I use the terms editing out a
 typo.  In my mind when you first write a document if you do
 something wrong that's a typo and when you proofread it, it
 becomes the editing error.  So I don't differentiate because I
 think I am not a -- scholar.

 Q  So mistakes in proofreading are editing errors and mistakes
 in typing are typographical errors.

 A.  I can't comment that because I do not feel that has anything
 to do with the science.

Tr. at 859.  Since ORI did not prove that Dr. Sharma understood and
adopted the position set forth in the letter from his counsel, and given
Dr. Sharma's manifest confusion when presented with this letter on the
stand, we are not able to conclude that the assertion that the
misstatement was caused by anticipating data and inadvertently failing
to correct it can fairly be attributed to Dr. Sharma.  The most
comprehensive written statement which Dr. Sharma made about his position
in no way suggested this analysis offered by counsel, and insisted that
the cause of the subscript substitution was typographical error.  Ex.
R-2.  We therefore do not give weight to the letter from Dr. Sharma's
counsel as evidence of intentional deception.

  c. Conclusion on inconsistent statements

Above, we have focused at length on the details of each purportedly
inconsistent statement by Dr. Sharma, and we have found them unavailing
to ORI's case.  We want to emphasize here, however, that, in our
attention to detail in this analysis, we have not overlooked the larger
picture.  We have considered the point of view that ORI's contentions,
even if not all substantiated individually, might cumulatively support a
finding that Dr. Sharma intentionally included erroneous information in
the first application.   We reiterate below some of the factors
discussed in more depth above which underlie our conclusion that ORI has
not shown a cumulative pattern supporting intentional deception.

As we discussed initially, the single subscript error involved here is
far more easily explained as inadvertent.  Nothing in the structure of
the grant application, its typing or preparation, or its perception by
the grant reviewers makes intentional deception a likely scenario.

From the beginning, when confronted with the error, Dr. Sharma
consistently stated in every interview that he was not aware of the
error when he submitted the application and that he believed it to be
typographical.  However, he made a variety of statements, in the course
of the long investigation, which referred in some way to his having
hoped to get sequence data on à2GC.  None of them were in the nature of
admissions of consciously inserting the wrong subscript.  Thus, we have
found:

 o       Those statements which Dr. Sharma made about having
 "meant to remove it," which ORI interpreted to mean that he must
 have known of the error in the first grant application, were in
 context responses about the revised application.  It is
 uncontested that the whole sentence was left in the revised
 application inadvertently and that Dr. Sharma did mean to take
 it out because it was completely unhelpful to his changed
 methodology and because he took out all the related sentences.

 o       Whenever asked about whether he was expecting or hoping
 to get sequence data on à2GC, Dr. Sharma consistently answered
 that he was, and evidence in the record supports the accuracy of
 this answer.  He was hoping to obtain this data at some point,
 but in order to conduct further experiments not to repeat the
 reported experiments using another probe.

 o       Since both categories of statements above can be
 understood as simply accurate responses to the questions asked,
 their repetition throughout the investigation is in no way
 suspicious.

 o       Dr. Sharma also speculated that he may have made the
 error for the very reason that the à2GC sequence data was so
 much on his mind, which does not imply any evil intent.

 o       Finally, this is not a case in which Dr Sharma claimed
 to have data which he had not yet received because he hoped to
 obtain it in time to justify the assertion.  The disputed
 sentence reports results of an experiment with a probe based on
 other data and receipt of the à2GC data would not have changed
 the inaccuracy of the subscript.

Therefore, we conclude that the statements attributed to Dr. Sharma do
not, individually or collectively, constitute persuasive evidence of an
intent to mislead.

 2. None of ORI's Other Arguments Demonstrate an Intent to
 Deceive

  a. Failure to provide a citation reference

At times, ORI suggested that Dr. Sharma's failure to cite the Lefkowitz
article after the disputed statement at page 21 in Exhibit H-2 in itself
constituted scientific misconduct.  See, e.g., ORI Final Report at
24-25.  At other times, ORI referred to the omission of such a citation
as additional evidence of an intent to deceive.  We do not find support
in the record for either interpretation.

At numerous other points in the same grant application, Dr. Sharma
plainly credited the Lefkowitz paper as the source of amino acid
sequence data on human platelet à2A receptor data.  Thus, it is not
plausible that Dr. Sharma was attempting to take improper credit for
Lefkowitz' work.  In retrospect, Dr. Sharma agreed that he probably
should have included a cite to the Lefkowitz paper after the disputed
language.  Tr. at 823.  However, he also explained at the hearing that
he had assumed that his reliance on the Lefkowitz paper would be obvious
from a mention of à2A because he had previously cited the paper in the
same section of the grant application and had described the work of the
Lefkowitz group sequencing à2A.  Tr. at 824; Ex. H-2, at 21.  Thus, he
stated that "since I did not know that this mistake had occurred, I did
not feel that I need to put the reference because everybody would have
identified that I could have only get, I could have only gotten the
alpha 2A probe which referred to this particular gene.  There was no
other way."  Tr. at 824.

Furthermore, ORI provided no testimony that the failure to cite
Lefkowitz again after the statement at issue would in itself have
constituted scientific misconduct, if the probe had been correctly
identified as corresponding to à2A data.  ORI's scientific expert
testified that he would cite more repetitiously, "ten sentences in a
row" if appropriate, but also stated that a scientist who was not so
careful to "overreference" would not be less ethical.  Tr. at 612-13
(Dr. Douglass).  We find that ORI did not prove that the failure to cite
to the Lefkowitz paper directly after the disputed statement on page 21
constituted scientific misconduct.

Nevertheless, we recognize that, had Dr. Sharma placed a cite to the
Lefkowitz paper at the end of the disputed statement, such a reference
would have undercut any claim that the subscript error was intended to
mislead.  This fact is demonstrated by the fact that ORI did not find
any scientific misconduct in Dr. Chalberg's identical error on page 34
of Exhibit H-2 in a description of Lefkowitz' work which stated that
"[r]ecently the cloning and resultant de novo expression of the à2GC
receptor in Xenopus laevis oocytes has been reported (43)."  The number
(43) is a cite to the Lefkowitz paper and the reference to à2GC is an
obvious error, since Lefkowitz's work was on à2A, and there was no
motive to falsely claim otherwise.  See ORI Final Report at 26; Tr. at
619-620.

While the omission of the citation was regrettable, we find that ORI did
not prove that it was evidence of intent to mislead or that it
constituted a serious deviation from commonly accepted scientific
practice.

  b. The pattern of misstatements

ORI suggested that it was more likely that the errors in Exhibit H-2
were intentional because the application contained "repetitive
misstatements."  ORI Final Report at 23.  However, the evidence before
us suggests that the other statements referenced in the allegations
relating to Exhibit H-2 were not independently false. 36/  Rather, they
were misleading only because a reviewer would refer back to the initial
misstatement (at Exhibit H-2 at 21) and conclude that they were
referencing à2GC amino acid sequence data which was not in fact
available.  See, e.g., Tr. at 556-60, 588-91, 631-32 (Dr. Douglass).
The ORI Final Report also noted that the other statements "refer back"
to the initial statement and do not themselves "specify which peptide
was used to construct the probe."  ORI Final Report at 23 and 30.
Nothing in the record demonstrated that these statements would be
problematic if the initial statement had correctly reported the probe as
based on à2A sequence data obtained from the Lefkowitz paper.
Therefore, it is unreasonable to rely on these statements as
constituting a pattern demonstrating the intentionality of the original
error on page 21, when the other statements would be accurate but for
that error.

In addition to finding a pattern of misstatements in the application
itself, ORI also offered evidence of a statement in a memorandum
apparently prepared for a retreat held in February 1990 (over a year
after the application was submitted) by the CCF Department of which Dr.
Sharma was a member.  Exhibit H-15.  ORI contended that this memorandum
showed another example of Dr. Sharma falsely claiming to have the à2GC
amino acid sequence data and was therefore "instructive."  ORI Final
Report at 24.  No testimony established who drafted or reviewed specific
portions of the summary or who actually received it.  Generally, the
memorandum appears to summarize in broad strokes the staffing, funding,
space and research directions and accomplishments of Dr. Sharma's
laboratory for the department as a whole.  ORI did not explain what
motive Dr. Sharma would have had to attempt to mislead members of his
own laboratory and department about the status of his research in this
document.  Members of the department not expert in this sub-specialty of
molecular biology would be unlikely to recognize the significance of the
specific claims; those most likely to recognize the significance were
members of his own laboratory who would also be most likely to catch any
inaccurate representations and many of whom were already on bad terms
with him by that date.  No evidence in the record suggested that this
statement on his research directions formed the basis of any benefits
which Dr. Sharma could have anticipated.

Furthermore, ORI did not prove that any statement in the summary was in
fact false.  Much of the language describing research directions is
almost identical to unchallenged language from the abstracts in the
corresponding grant applications.  Compare Exhibit 15, at 2-3 (on
à2-adrenergic receptor) with Exhibits H-2 and H-35, at 2; compare
Exhibit H-15, at 3 (on atrial natriuretic factor receptor) with Exhibit
H-3, at 2.  The summary also contains an assertion, apparently the one
contested by ORI, that cDNA has been cloned for an à2GC receptor subtype
which is "brain-specific," and that for this receptor the nucleotide
sequence of this clone has been determined and from it the amino acid
sequence has been inferred.  Exhibit H-15, at 3.  Since the summary
indicates that this member of the à2-receptor family does not appear in
the adrenal gland, the work referenced obviously here related to the
research proposed in the revised application (Exhibit H-35), not to
Exhibit H-2.  In fact, the questioned statement is almost identical to
one in Exhibit H-35 which asserts that Dr. Sharma's laboratory had "now
cloned, sequenced, and expressed a rat brain à2-adrenergic receptor
cDNA."  Exhibit H-35, at 21.  A listing of the 31 amino acids from a
peptide which has been synthesized follows on the same page.  Id.  ORI
found no scientific misconduct after its review of Exhibit H-35.  ORI
Final Report at 31-33, 54.  Therefore, to the extent that this summary
simply restates the unchallenged status of research set forth in Exhibit
H-35, we do not see any basis in the record to conclude that the summary
evidences a pattern of intentional misstatements.

  c.  The failure to report to NIH

ORI argued that Dr. Sharma's failure to report his misstatements to NIH
evidenced "knowing deceit" and was "a deliberate attempt to prevent
detection." 37/  ORI Post-Hearing Br. at 30.  ORI's scientific expert
testified that whether he would notify a funding agency of an error in a
grant application would depend on the significance of the mistake in
context and that there is "a huge spectrum with regards to errors and
what the type of error you would notify the agency" about.  Tr. at
584-5.  On the other hand, when asked if an error "of the magnitude" of
that identified by ORI in Exhibit H-2 was "worthy of notification to the
funding agency," he responded that he "would think so." 38/

However, it begs the question to use the failure to contact NIH before
the allegations arose as proof of bad intent, since obviously not
reporting an error is just as consistent with not being aware of its
existence.  Therefore, we must look at whether Dr. Sharma's actions
after learning of the misstatements were inappropriate.  The earliest
point at which anyone has testified that Dr. Sharma's attention was
directed to misstatements in his grant applications was in April 1990.
Within two weeks of the initial allegations, the Healy Committee had
begun the formal investigation process.

By that time, both grant applications at issue here had been rejected
for funding.  It is difficult to see what benefit NIH would gain from a
notification that an error had existed in a grant application which it
had already determined not to fund.

As to the revised, funded grant application, Dr. Sharma argued that he
did not report the misstatement which remained because it was already
the subject of an investigation, which he assumed would be made known to
NIH.  Tr. at 847, 852.  In addition, even ORI has concluded that the
statement in the revised application no longer was material to the
revised methodology, so we have no basis to conclude that the error in
that grant application was so significant that notification was clearly
demanded by the circumstances.

ORI also characterized Dr. Sharma as having "resisted informing NIH that
his revised application contained false statements."  ORI Post-Hearing
Br. at 30.  This position is based on two documents.  The first was
submitted to ORI by Dr. Sharma which stated that "[a]fter the committee
deliberations, extraordinary pressure was put on me to admit to being
guilty.  Dr. Hamilton asked me to write to NIH, requesting surrender of
my grant."  Ex. R-2, Part II, at 3.  The context indicates that this
conversation occurred in early September 1990.  The second document was
a memorandum dated September 11, 1990 from Dr. Sharma to Drs. Healy and
Hamilton, apparently in response, stating:  "I respectfully request your
consideration to read my response to the second review committee.  I
will then follow your instructions on the submitting of a letter to NIH
regarding my recently funded grant."  Ex. H-10, at 1.  The action
requested by Dr. Hamilton was more drastic than informing NIH of a
misstatement.  Dr. Sharma's response amounted to agreeing to submit to
the joint decision of Drs. Healy and  Hamilton as to what letter to send
to NIH, with the one request that they read his comments before
deciding.  His attached comments again explained that the original grant
application contained an unintentional typographical error and then
pleaded that "[t]his seems to be a slim reed indeed upon which to
destroy a dedicated and accomplished researcher's career."  Ex. H-10,
Attachment 1, at 1-2.  Such a statement can hardly be read as a refusal
to inform NIH of a misstatement.

In a related point, ORI argued that Dr. Sharma's persistent focus on his
view that confidentiality had been breached in the distribution of his
grant applications demonstrated that he was "more concerned that the
errors had been detected and how they had been detected than the fact
that there were errors in the applications."  ORI Post-Hearing Br. at
30-31.  Undoubtedly, Dr. Sharma can be seen in many documents to have
focussed on his concern about how Dr. Chalberg happened to have obtained
applications for research he was not involved with, in light of the
thick atmosphere of suspicion and conflict that appears to have existed,
as discussed above in relation to the statements reported by Drs.
Chalberg and Ferrario.  See Ex. R-2, at 8.  However, we do not see the
connection that ORI makes from Dr. Sharma's feeling of persecution to a
"deliberate attempt to prevent detection of his overstatements."  ORI
Post-Hearing Br. at 30.  The misstatements had already come to light at
the time that he expressed the complaints about distribution of his
grants to which ORI referred, so he could hardly have hoped to somehow
prevent their detection.  Furthermore, Dr. Sharma's concern about broad
distribution of his grant applications has not been proven to be
entirely unfounded in what ORI described as the "competitive realm of
grant applications."  ORI Post-Hearing Br. at 30.

  d. ORI's other attacks on Dr. Sharma's credibility

ORI questioned Dr. Sharma's overall credibility on the basis that he
denied that he was ever called to a meeting on April 20, 1990 with Dr.
Ferrario and Dr. Bannerjee to answer Dr. Chalberg's charges of
scientific misconduct and that he denounced Dr. Ferrario's memorandum of
that meeting as a "fabrication."  Tr. at 22; ORI Post-Hearing Br. at
32-24; see also Ex. H-23, at 2-3.  While Dr. Sharma did say the meeting
as reportedly described by Dr. Ferrario never occurred and the
memorandum is a fabrication, he also said in the same ORI interview that
he did meet with Mr. Murray and Dr. Ferrario one day between April 10th
and April 20th.  He appeared to base his refusal to agree to the
specific date largely on the fact that he found no record of a meeting
on that date in his personal diary.  See also Ex. H-23, at 2.  However,
it is clear in the rest of the interview that he recalled that a meeting
had occurred.  What he most violently objected to was the
characterization of this meeting as one where Mr. Murray called him in
to answer Dr. Chalberg's charges.  Rather, he described feeling already
embattled in the department when he learned from a secretary that Dr.
Chalberg was obtaining copies of his grant applications and raising
questions about them. 39/  He then described the meeting as a brief one
40/ in which he complained to Mr. Murray and Dr. Ferrario about why Dr.
Chalberg was being given free access to his grants.  Dr. Sharma
acknowledged that he was informed that charges of scientific misconduct
had been raised against him, but consistently denied that he was
provided with the specific charges relating to the misstatements about
à2GC before the Healy Committee (and that even then the applications at
issue remained unclear).  Dr. Sharma said that he did not recall Dr.
Bannerjee being present at a meeting with Dr. Ferrario "where they
discussed the alpha 2GC receptor" with him or where they discussed
"general problems with any grants or problems in the lab."  Ex. H-31, at
363.

We find that credible testimony supports the conclusion that a meeting
occurred on or around April 20, 1993 in which Dr. Ferrario communicated
to Dr. Sharma that some questions about the integrity of his grant
applications were being raised and Dr. Sharma communicated that he was
upset about the apparent distribution of his grant applications.
However, we do not find the inconsistencies in the various participants'
memory of this meeting (including significant differences which have
been noted between Dr. Bannerjee's and Dr. Ferrario's accounts)
sufficient to demonstrate dishonesty on the part of Dr. Sharma.

ORI further challenged Dr. Sharma's credibility based on a letter he
sent to ORI during the investigation, in which he suggested that a
number of scientific points would have made the error in Exhibit H-2
obvious and therefore would serve to disprove the allegation of
intentional deception.  Ex. R-2.  Dr. Sharma did not press these points
before the Panel.  However, ORI alluded to these arguments as evidence
that Dr. Sharma was "merely grasping at straws" in seeking to explain
the misstatement in his grant application.  ORI Post-Hearing Br. at 38.

The first point Dr. Sharma made was that the reference to use of rat
codon bias data would have alerted reviewers that he was translating
from another species and therefore supported his intent to refer to à2A
(human platelet) data instead of à2GC (rat) data.  Ex. R-2, at 2.  ORI
argued to the contrary that rat codon bias data would be consulted
regardless of the source of sequence data for the probe.  ORI
Post-Hearing Br. at 38.  Elsewhere, ORI went further and asserted that
both human and rat codon bias data would have been referenced if Dr.
Sharma were, as he claimed, intending to refer to translating across
species (i.e., from à2A to probe rat cells).  ORI Post-Hearing Br. at
16.  The only evidence to which ORI cited to support this assertion is
testimony by Dr. Chalberg that is inconsistent with it.  Tr. at 164.
Dr. Chalberg testified that "you would only look at a rat bias table if
you were trying to make a rat-specific probe, which we were trying to
do, based upon human sequence."  Id.  Dr. Douglass also indicated that
in using data from one species to construct a probe to search for
clones, whether in the same or another species, one uses bias data for
the target species. 41/  The weight of the scientific evidence is that
the reference to use of rat codon bias data in construction of the probe
does not offer any conclusive information about the intended source of
the probe.  See, e.g., Tr. at 167 (Dr. Chalberg), 571-2 (Dr. Douglass).

We find that Dr. Sharma's other points do not prove that the error was
so obvious as to demonstrate his innocent intent.  For example, Dr.
Sharma also suggested that his use of 90% stringency conditions would
have alerted researchers that he was seeking to hybridize to a probe
based on sequence from something other than the target protein.  Ex.
R-2, at 2.  However, the scientific evidence suggests that stringency
conditions cannot meaningfully be defined with sufficient mathematical
precision to make this point obvious to a reviewer.  Tr. at 598-9,
627-30.  Similarly, Dr. Sharma suggested that the 39-oligomer length of
the probes would have provided reviewers a clue that he had used Dr.
Lefkowitz's à2A probes, since they were identical in length.  While the
probes were the same length, since Dr. Sharma did, as we have found, use
the à2A probes, the evidence established that that length was not an
uncommon one for such probes and that reviewers would not have been
likely to make a connection between that length and a probable reliance
on the Lefkowitz probes.  See Tr. at 230, 582.

For the reasons explained above, we do not rely on these points in our
determination that the error was not intentional.  However, we also find
that ORI did not prove that these arguments were so implausible as to
undercut Dr. Sharma's credibility for having raised them.  e. Summary on
  ORI's other arguments on intent

Overall, we find that the collateral arguments offered by ORI to
buttress the allegation of intentional deception are largely based on
hindsight.  While the error here might have been prevented by such means
as more frequent citations, more careful proofreading, or better review
procedures, ORI did not show that Dr. Sharma's practices in these areas
deviated from the normal range of practices of scientists at that time.
There is thus no basis in these criticisms to conclude that his failure
to catch the error means that he intentionally inserted it.  Finally, we
are not persuaded that the disagreements about the April 20, 1990
meeting or the points made by Dr. Sharma in an effort to explain why an
inadvertent error was more consistent with the rest of the sentence
undermine Dr. Sharma's credibility.

 D. Conclusion Regarding First Grant Application

For all the reasons discussed above, we conclude that ORI has failed to
prove by a preponderance of the evidence that Dr. Sharma included any
misstatement in Exhibit H-2 intentionally to mislead reviewers.

II.  The Second Grant Application

The charge relating to Exhibit H-3 was based on a single clause,
highlighted below, in one sentence on page 23, which read as follows:

 [C]omparison of our unpublished sequence data of the 180-kD
 protein with that of published rat brain protein indicate
 significant structural heterogeneity between certain
 CNBr-cleaved fragments of the two proteins.

Dr. Sharma admitted that he did not have unpublished amino acid sequence
data on the 180 kD protein, but explained that he meant to write
unpublished peptide map data from the sequencing process.  Ex. H-30, at
125; Respondent's Post-Hearing Br. at 5-6.  As explained above, amino
acid sequence data generally means the precise order of amino acids in
either some region of or the entire length of a protein.  Peptide map
data results from an intermediate step in the process of determining the
sequence of a protein which is based on cleaving the protein by
chemicals which reliably divide the chain in predictable ways.  The
information available from peptide mapping was described by ORI as
"points of the sequence."  ORI Final Report at 43.  ORI acknowledged
that "a peptide map is a type of amino acid sequence data."  Id.  It is
not disputed, however, that possession of full amino acid sequence data
would indisputably represent a more advanced step in the research
process than possession of peptide map data alone.  Tr. at 493, 636.

Dr. Ravi Marala, who worked on this project with Dr. Sharma at CCF, had,
by June of 1989 when this application was submitted, been working toward
obtaining amino acid sequence data for "quite a while," and had obtained
peptide map and isolated fragments which were sent to Dr. William
Merrick for sequencing without success.  Tr. at 490-2.  On this basis,
ORI concluded that Dr. Sharma was portraying his research as
substantially more advanced than it was.  ORI Post-Hearing Br. at 25.

As mentioned above, ORI placed far less weight on this allegation than
on the charges relating to à2GC, as evidenced by the fact that only two
of the 44 pages in ORI's post-hearing brief were devoted to the 180 kD
protein issue.  In large part, ORI treated this allegation as resting on
a continuation of its claim that Dr. Sharma was exhibiting a pattern of
falsely claiming to be further along in sequencing proteins than he was.
See ORI Post-Hearing Br. at 26.  Since we have found ORI's basis for
charging this pattern unsubstantiated with regard to Exhibit H-2 above,
we do not find that it provides any foundation for the allegations with
regard to Exhibit H-3.

ORI again attempted to demonstrate that Dr. Sharma had admitted that he
had referred to sequence data because he anticipated obtaining the data
in order to justify the reference in time or removing the reference
before submitting the grant.  ORI Post-Hearing Br. at 25, n.63.  The
only statement on which ORI relied for this purpose was made by Dr.
Sharma to the Hamilton Committee:

 DR. HAMILTON:   . . . When you say that the use of sequence per
 se was erroneous, that you really intended or should have said
 peptide map data, when you wrote this was it written in an
 anticipatory manner --

 DR. SHARMA:  That's right.

 DR. HAMILTON:  -- that is, you expected you would have that
 sequence --

 DR. SHARMA:  That's right, that's right.

Ex. H-29, at 16.  Dr. Sharma testified that the intention of his
statement was to agree that he intended to write peptide map data and to
agree that he was anticipating (was in fact actively in the process of
seeking) sequence data on the 180 kD protein at the time of the grant.
Tr. at 861-3.  However, he testified that the statement "did not mean
that I was getting it to put it in the grant."  Tr. at 863.

As with the à2GC sequence data, so in relation to 180 kD protein
sequence data, Dr. Sharma agreed with many questioners that he was
anticipating getting sequence data on peptides of 180 kD, and in fact he
later got some of that data.  See, e.g., Ex. H-31, at 353-8; Tr. at 257.
However, he consistently insisted that these responses did not imply
that he put in the reference to sequence data intentionally in
anticipation that he would either obtain data to justify before the
application was reviewed or would remove the reference.  Id.  He
asserted throughout that he meant to refer to peptide map data.  Id.

The most telling argument that Dr. Sharma did not insert the reference
to sequence data in Exhibit H-3 with the intent to mislead reviewers is
the fact that the application contained a number of statements which
appeared to contradict, more or less directly, the claim to have such
data.  As a result, the critique from the study section evaluating the
application pointed out some of the contradictions and asked for
clarification.  Thus, the reviewers asked:

 On page 23, it is stated:  "...comparison of our unpublished rat
 brain protein indicate significant structural heterogeneity
 between certain CNBr-cleaved fragments of the two proteins.",
 while on page 34: "...(oligodeoxynucleotide) probes will be
 synthesized as soon as sufficient sequence data on the guanylate
 cyclase protein is available.", and finally on page 35:  ...we
 presently have approximately 3 nmol of purified 180-kD for
 N-terminal amino acid sequencing.".  Which of the above is true?
 If the amino acid data are available, why aren't they shown?

Ex. R-8, at 3 (emphasis added).  Plainly, the reviewers were not misled
and the language of the grant was so puzzling on its face that it was
highly unlikely to mislead. 42/  While ORI argued that the failure to
deceive the reviewers does not "vitiate the impropriety of Dr. Sharma's
attempt," the idea of attempting to mislead someone by telling them two
or more contradictory, or at least apparently inconsistent, things is
improbable.

Dr. Douglass suggested that the two statements could be reconciled to
mean that an inadequate amount or quality of amino acid sequence data
was available ("maybe they've only got five or seven amino acids and
that's not enough"), so they are waiting to have sufficient amino acid
sequence data.  Tr. at 635.  However, he concluded that "there just
isn't enough information as to what they do have to make any type of
correlative comments with regard to these two sentences."  Tr. at 636.
His suggestion would not answer the reviewers' second question, however,
i.e., why no data is presented if any is available.

In relation to the inaccurate description of the data which Dr. Sharma
had, we note that (1) the term sequence data is potentially ambiguous,
and (2) Dr. Sharma was not expert in the use of technical molecular
biology terminology.  On the first point, Dr. Marala pointed out that
the term sequence data could be ambiguous outside of a known context,
and that therefore one should "be more specific if . . . talking to a
third person who is not involved in that research."  Tr. at 509.

The imprecision of the term sequence data was compounded by Dr. Sharma's
frequent use of inexact terminology, in contexts where no motive to
deceive was suggested.  See, e.g., ORI Final Report, Tab S at 1 (Letter
from Dr. Bylund to OSIR).  Thus, ORI reported that its scientific
advisors noted that Dr. Sharma's training was in medicinal chemistry
rather than molecular biology, and that, perhaps as a result, much of
the technical writing in his protocols is "phrased awkwardly and in a
manner inconsistent with standard terminology."  ORI Final Report at 27
(noting, as we have above, that Dr. Sharma's "conversational speech also
is characterized by idiosyncratic phrasing").

This particular grant application received an extremely poor evaluation
by reviewers.  ORI Final Report at 41.  Dr. Sharma repeatedly expressed
during the investigation his embarrassment about the low quality of
writing in this application and stated that he wrote it very hurriedly.
See, e.g., Ex. H-30, at 126-8 (OSIR Interview of Dr. Sharma).  ORI
acknowledged in its final report that the "poor quality of [Exhibit H-3]
. . . , characterized by obvious inconsistent statements, suggests that
Dr. Sharma lacked the necessary intent to deceive or mislead and that
many misstatements resulted from hasty preparation . . . [and] his lack
of expertise in molecular biology."  ORI Final Report at 42.  ORI
nevertheless found scientific misconduct, because it found the statement
at issue analogous to the misstatement in Exhibit H-2 and because it
found that Dr. Sharma was aware that peptide map data was less advanced
than most scientists would understand by a reference to sequence data.
Id. at 43.  We find that these arguments unsupported for the reasons
discussed above.

We find that the misstatement in Exhibit H-3 was the result of careless
error and that ORI did not prove that Dr. Sharma intended to mislead
reviewers about the state of his research on 180 kD protein.  Therefore,
we conclude that ORI has failed to prove by a preponderance of the
evidence that Dr. Sharma included any misstatement in Exhibit H-3
intentionally to mislead reviewers.

         Conclusion

For the reasons explained above, we conclude that ORI did not prove by a
preponderance of the evidence that Dr. Sharma committed scientific
misconduct.  Consequently, we conclude that ORI's findings are not
supported and the proposed administrative actions are not justified.

       ___________________________
       Judith A.
       Ballard

       ___________________________
       Norval D. (John)
       Settle

       ___________________________
       Donald F.
       Garrett
       Presiding Panel
       Member

1.   The following discussion is based on the record as a whole before
the Panel.  Disputed scientific questions are discussed in the text
where necessary.  The Panel permitted the parties to submit proposed
stipulations relating to scientific terminology.  ORI submitted a set of
definitions of scientific terms on which we have relied to the extent
that they were not objected to or were supported by the record.

2.   A revised version of this grant application (Exhibit H-35), with a
dramatically altered research methodology, was submitted to NIH in
October 1989.  The revised application did receive funding.  The
revised, funded grant is not at issue before us, but allegations were
raised earlier in the investigations about both Exhibits H-2 and H-35.
At times, the two substantially different versions of the à2GC research
grant applications created confusion for investigators and witnesses.

3.   Allegations 2 and 3, which arose during earlier stages of the
investigation, as well as other subparts of Allegations 1 and 4, were
determined by ORI not to constitute scientific misconduct and are
therefore not at issue before us and are not quoted in the text.  See
ORI Final Report at 54-55.  These other misstatements were found to be
"caused by careless yet honest errors - typographical, editing, and
supervisory."  Id. at 55.

4.   The Ruling as a whole is not repeated here, but is incorporated by
reference.

5.   There is no doubt that Dr. Sharma knew he had a duty to be accurate
in his grant applications.  A principal investigator must certify in a
grant application (as Dr. Sharma did by his signature on these
applications) that he is aware that "[w]illful provision of false
information is a criminal offense."  However, this certification by
itself does not support a conclusion that the investigator would be
subject to criminal or even administrative actions for unintentional or
even negligent errors alone.

6.   Dr. Sharma told ORI that he gave Exhibit H-2 to Dr. Sen and Exhibit
H-3 to Dr. Bannerjee before submission, although it is not clear that
either one actually did review the applications.  Ex. H-30, at 129; Ex.
H-31, at 370-3 (Dr. Sharma's OSIR Interviews).  Each application was
also certified by the Director of CCF, as well as by Dr. Sharma.  Exs.
H-2, H-3.

7.   In its discussion of negligence, ORI also argued that Dr. Sharma
could have prevented the error by citing Lefkowitz directly after the
statement, that he had a pattern of errors demonstrating disregard for
the accuracy of his grant applications, and that his failure to take
action to notify NIH of the error demonstrated disregard for accuracy.
ORI Post-Hearing Br. at 28-31.  All of these points were also cited by
ORI as arguments in support of its claim of intentional deception.  See
id. at 30.  We discuss the evidence relating to each of these areas in
more detail in relation to the analysis of intentional deception.
However, we note here that ORI did not prove that the failure to add an
additional cite to Lefkowitz at that point was a serious deviation from
accepted citation practices, that the errors which were not raised as
allegations before us formed a pattern evidencing a serious deviation
from commonly accepted practices of grant preparation, or that any
commonly accepted practice required notifying NIH of an unintentional
error in a grant application which had been rejected for funding and
which was already the subject of an investigation.

8.   The other two disputed statements appeared on pages 27 and 32 of
Exhibit H-2.  Each contains a reference to 39-mer probes "corresponding
to the amino acid sequences of two internally-cleaved peptides."
Neither statement asserts that the probes were based on à2GC data.
However, a reviewer would assume that these statements refer back to the
39-oligomer probe first described on page 21 of the application.  The
erroneous reference to à2GC instead of à2A  probes on page 21 therefore
had the effect of making the other two statements false.  See ORI Final
Report at 23.

9.   The weight of scientific testimony at the hearing was that
scientists can use a probe based on a known protein sequence of a member
of the same protein "family," which here was the family of à2 adrenergic
receptors, in order to look for related proteins.  See, e.g., Tr. at
208, 215-6 (Dr. Marshak).  Thus, it was not an unreasonable scientific
strategy for Dr. Sharma to have used the published data on à2A to
construct a probe in order to search in a different set of cells for a
related protein, tentatively called à2GC.  See Tr. at 587 (Dr.
Douglass).  His argument for the existence of such a related, but not
yet conclusively identified, protein rested on his studies of its
biochemical and pharmacological properties, which differed in important
ways from à2A.

10.   ORI speculated in its post-hearing brief that substituting à2A for
à2GC would not "cure" the inaccuracy in the sentence, because a
researcher having access to published sequence data "would" have
constructed a genomic probe rather than using the 39-oligomer probes.
ORI Post-Hearing Br. at 14.  ORI based this speculation on an entire
textbook (Exhibit H-38) which does not directly support the point made.
ORI cited no testimony whatsoever that Dr. Sharma's laboratory had not
in fact followed the described procedure.  Dr. Chalberg was asked at the
hearing whether the disputed statement would "be an accurate statement,"
with the single substitution of à2A for à2GC, and responded:  "[T]hat's
what we actually did do."  Tr. at 124-25; see also Tr. at 880-1 (Dr.
Duda, who performed the actual experiments, testifying they got
39-oligomer probes based on à2A sequence data from Dr. Michael Cashnell
at NIH, and used them to screen the cDNA library as described in the
disputed sentence).  ORI's scientific consultant conceded that the
sentence in context would make sense and be accurate if the highlighted
reference to à2GC had been replaced by a reference to à2A.  Tr. at
586-7, 631-2 (Dr. Douglass); see also Tr. at 125-7 (Dr. Chalberg).  ORI
was certainly on notice that this issue was critical and that ORI bore
the burden of proof, yet ORI offered no testimony to support its
speculation.  For example, ORI did not offer an affidavit or testimony
from Dr. Cashnell at NIH denying that he synthesized a 39-oligomer probe
using the à2A data.  Instead, ORI cited a comment by Dr. Sharma that
screening with a 39-oligomer probe, instead of a cDNA probe, "would have
been a folly," which ORI treated as an acknowledgement that Dr. Sharma
did not use a 39-oligomer probe in the work reported in Exhibit H-2.
Ex. R-2, at 3; ORI Post-Hearing Br. at 14, n.31.  This interpretation is
wholly unwarranted.  In context, Dr. Sharma was clearly discussing why
retaining the reference to a 39-oligomer probe in his revised grant
application (Exhibit H-35) would have been a "folly" because his
methodology there was different and relied on a cDNA clone as a probe.
Scientific testimony at the hearing indicated that, while a longer probe
might often be preferable, many factors go into the selection of a probe
for a particular research purpose.  See, e.g., Tr. at 539, 608-12.
Thus, we conclude that the statement in Exhibit H-2 was accurate but for
the substitution of the subscript.

11.   However, it is not clear that the use of an à2GC probe would
necessarily have been more likely to succeed in hybridizing to clones of
à2GC.  The two related proteins were estimated by Dr. Sharma to be over
90% homologous, i.e., to have more than 90% identical amino acid
sequences.  Tr. at 865; see also Tr. at 875-9 (Dr. Duda testifying that
the two proteins were "almost identical," except for their
distinguishing chemical behavior).  To the extent that the identical (or
"conserved") portions of the amino acid sequence were used in designing
the probe, it would be impossible to know which protein was represented
by a clone hybridized to the probe, regardless of whether the probe
corresponded to à2GC or à2A, except by using the clone to produce the
protein itself and testing the behavior of the protein ("expression") or
by sequencing the clone and comparing it to known sequence data.  Tr. at
955-6 (Dr. Gelmann).

12.   Dr. Sharma made the same point in comments he submitted to Dr.
Lawrence J. Rhoades, Ph.D., at OSIR, in which he stated that "nowhere in
the application is it indicated that we had the sequence of the
à2GC-fragments.  If we had that information, we would have identified
the sequence and described it . . . ."  Ex. R-2, at 2; see also Ex.
H-29, at 55-57.

13.   ORI cited to no evidence contradicting Dr. Gelmann's assertions
about the level of detail a reviewer would have expected to see if amino
acid sequencing data had been obtained, but argued that Dr. Gelmann's
testimony should be "discounted substantially" for five reasons: (1) he
"never purified a receptor membrane protein;" (2) he did not have a
doctoral degree; (3) he "never worked with adrenergic receptors;" (4) he
was not familiar with literature and reported research; and (5) "he had
not reviewed the entire grant applications" at issue here.  ORI
Post-Hearing Br. at 13, n.30.  We find Dr. Gelmann's testimony credible
and decline to discount it on the bases suggested by ORI.  Dr. Gelmann
has an M.D. from Stanford University, rather than a Ph.D.  Tr. at 902;
Ex. R-15, at 1.  ORI did not demonstrate that a Ph.D. was required in
the field.  Dr. Gelmann did research on tumor cell biology at the
National Cancer Institute and was a tenured scientist at NIH.  Tr. at
902; Ex. R-15.  He "has been involved in DNA cloning since 1979,"
receives NIH grants to do gene cloning, and clearly has expertise in
molecular biology.  Tr. at 903-4, 911.  Furthermore, he has been a full
member of one study section and an ad hoc member of another, as well as
sitting on other scientific review committees.  Tr. at 903; Ex. R-15, at
3.  He testified that he did not "have difficulty reviewing the science"
in the applications, "[i]n particular as it regarded DNA cloning and
some general aspects of protein sequencing."  Tr. at 940.  Furthermore,
he testified that he did have "some familiarity" with the biochemistry
and pharmacology of adrenergic receptors, even though he had not himself
cloned a membrane receptor protein.  Tr. at 911, 940.  Although there
was some testimony that reviewers should be familiar with the relevant
literature, ORI did not show why such familiarity with the literature on
a particular protein was needed to explain what reviewers would expect
to see in a grant application announcing sequencing of a new protein.
See ORI Post-Hearing Br. at 13, n.30; Tr. at 221, 940.  Dr. Gelmann had
not been able to review the entire applications, but testified that he
was able to draw conclusions about the scientific questions based on the
portions which he had reviewed.  Tr. at 929, 940.

14.   ORI denied that reviewers would expect to see the actual data in
the application, but this claim is undercut, in addition to the
testimony referenced above, by the reviewers' evaluation of Dr. Sharma's
180 kD protein grant application, Exhibit H-3.  In their evaluation, the
reviewers ask:  "If the amino acid data are available, why aren't they
shown?"  Ex. R-8, at 3.  This question, though not relating to Exhibit
H-2, strongly corroborates the testimony that reviewers expect to see
actual amino acid sequence data in grant applications when it is
available.

15.   In fact, the definitions of scientific terms submitted by ORI
support the conclusion that data less conclusive than amino acid
sequence can provide substantial information on the structure and unique
identity of a protein.  Thus, biochemical analyses can provide
information on the size, functional role, and cell location of a
protein.  ORI Scientific Terms at 1.  Peptide mapping can provide pieces
"unique enough to be used as a `fingerprint.'"  Id.  There is no dispute
that Dr. Sharma had biochemical and peptide data on à2GC.

16.   ORI also argued that the focus of the application on proving the
"uniqueness" of the à2GC receptor was significant in making the error
misleading to reviewers.  ORI Post-Hearing Br. at 15.  "Uniqueness" of
the à2GC receptor was, in context, a working hypothesis that a receptor
in rat adrenocortical carcinoma cells behaved differently from
previously known à2 receptors and was expected to have at least some
distinction in structure to account for this different behavior.
Exhibit H-2;  see Tr. at 864-5 (Dr. Sharma); Tr. at 875-9 (Dr. Duda).
This hypothesis had been supported by pharmacological, biochemical and
preliminary structural data.  Id., Tr. at 944-5.  Uniqueness is in no
way contradicted, despite ORI's assertion to the contrary, by Dr.
Sharma's expectation that the homology (i.e., shared amino acid
sequences) between à2GC and the known à2A  receptor was likely to be
very high.  Cf. ORI Post-Hearing Br. at 15, n.34, Tr. at 865.  Two
proteins can be very similar and yet differ in small but significant
ways.  Tr. at 868.  In order to prove the uniqueness of the à2GC, as
noted above, Dr. Sharma would ultimately have had to isolate clones
which proved on expression to represent à2GC and to then obtain full
amino acid sequence data.  See, e.g., Tr. at 561 (Dr. Douglass
testifying that proof of "uniqueness" will depend on full amino acid
sequence data); 821, 865 (Dr. Sharma).  The use of an à2GC probe in
itself would not have demonstrated the uniqueness of à2GC in any case.
Rather, that was a final aim of the project.

17.   The revised application is not the basis of an allegation of
scientific misconduct before us, since ORI concluded that the sentence
containing the reference to à2GC sequence data was left in that
application as a result of editing error.  ORI Final Report at 32-33.
As noted above, the revised application was dramatically different in
methodology.  In the revised application, Dr. Sharma used an à2A genomic
probe (rather than an à2A  39-oligomer probe) to isolate a cDNA clone
from rat brain cells (rather than rat carcinoma).  He was able to
sequence and express this cDNA, determine that it represented an à2
receptor, and then use it as a probe for further screening.  Exhibit
H-35; see generally ORI Final Report at Tab S at 3-5 (Letter from Dr.
David Bylund to OSIR).  In light of these changes, the disputed
statement, which had been carried over verbatim from Exhibit H-2, no
longer had any potential to benefit Dr. Sharma.  Throughout the
investigation, confusion persisted in interviews, when Dr. Sharma
referred to an intent to change this sentence or to leaving this
sentence over, about whether he meant he intended to change the
subscript in Exhibit H-2 if he did not get anticipated data in time to
justify it or whether he meant he intended to omit the sentence in
Exhibit H-35 because it was not relevant there but overlooked it in his
revision.

18.   When asked at the hearing how the error got in the sentence at
page 21 in Exhibit H-2, Dr. Sharma replied:  "I do not know that.  I do
not know to this date.  If I knew, it would not be there."  Tr. at 868.

19.   Dr. Sharma was asked in his first OSIR interview whether he wrote
Exhibit H-35 (the revised application) all by himself and responded that
he did, a statement which ORI also cited in support of its assertions in
the text.  Ex. H-30, at 9.  Aside from the fact that the question
related only to the revised application not before us, having written
something oneself plainly does not preclude receiving secretarial
assistance in typing and formatting it.  And in fact, in relation to
Exhibit H-2 (the application at issue), Dr. Sharma told ORI that "I
only, I typed it but I always get secondary help from my wife, you know,
she's, because I do not, I did not know much typing at that time.  And
my wife always helps me, volunteers to help me in typing."  Ex. H-31, at
218.

20.   ORI argued that Dr. Sharma would have noticed and had time to
catch the error if it were unintentional, because he began writing the
application in November 1988 and did submit until February 1989.  ORI
Post-Hearing Br. at 18; .  However, there is no evidence that he wrote
this section this section in advance.  In fact, the experimental results
reported in the sentence at issue were not completed until January 1989.
Tr. at 888-90 (Dr. Duda).  While ORI interpreted this timing to mean
that Dr. Sharma must have focussed on the sentence again at that time,
it is as likely that he wrote this section for the first time at that
late date.

21.   For example, ORI noted that it was difficult to determine what was
meant in Sharma's interview before the Hamilton Committee "because the
questions . . . asked were not precise and Dr. Sharma often interrupted
before the question had been asked and he had an opportunity to reflect
on his answer."  ORI Final Report at 22.

22.   For one thing, claiming to have cited non-existent data but to
have hoped to obtain it or make a correction is not exculpatory, yet Dr.
Sharma behaved as if he thought he had offered explanations of honest
error and expected exoneration rather than as if he had confessed.
Second, in most cases, Dr. Sharma also said something about mistyping or
missing the mistake during the course of the very same interview.  It
would be unreasonable for him to offer two mutually exclusive
explanations for the same error simultaneously, so it is more plausible
that he was trying to make different points.  Third, while Dr. Sharma
ultimately hoped to obtain à2GC sequence data, the evidence does not
show any reasonable basis for him to have expected its receipt during
the relevant time frame immediately before or after submission of
Exhibit H-2.

23.   ORI argued that Dr. Sharma must have said that he wrote à2GC in
Exhibit H-2 because he anticipated getting the data, since "[f]our
different groups of individuals" thought he had.  However, those groups
did not arrive at their understandings independently.  Rather, they had
each reviewed Dr. Ferrario's initial claim that Dr. Sharma had offered
that "explanation," as well as interviewing Drs. Chalberg, Ferrario and
Bannerjee.  In reading the later reports and transcripts, it is obvious
that the later reviewers were influenced by the initial impression,
right or wrong, which had been formed of Dr. Sharma's position, and they
framed their questions and heard his answers in that context.

24.   In its oral argument and post-hearing brief, ORI numbered the
various statements which it interpreted as multiple explanations.  For
example, ORI grouped certain statements Dr. Sharma allegedly made to Dr.
Chalberg (that he had gotten the data or that the statement did not
appear in the applications as submitted) as the first and second
explanations, considered a statement from Dr. Ferrario's report of his
meeting with Dr. Sharma (that Dr. Sharma included the statement because
he thought he would have the data before the application was reviewed)
the third explanation, and certain statements made to the Healy and
Hamilton Committees and in ORI interviews (to the effect that the
retention of à2GC was an editorial error by forgetting to remove the
anticipatory statement) the fourth explanation, and typographical error
the fifth explanation.  We find these groupings to be misleading and
have not used them.

25.   ORI also relied in support of the Healy Committee's interpretation
of Dr. Sharma's explanation on an anonymous committee member who
allegedly told ORI that Dr. Sharma said that "he `fully expected' to
have the sequence data and `that is why he wrote it in' the
application."  ORI Post-Hearing Br. at 35, n.102.  We give no credence
to this comment, since ORI presented the inconsistency of Dr. Sharma's
statements as central to its case and yet failed to produce this
witness.  ORI cited only to its own final report, which failed even to
identify the committee member.  ORI Final Report at 15.

26.   The transcript of the Hamilton Committee interview must be
discounted to some extent in light of several problems.  First, Dr.
Hamilton testified that there were gaps in the taping of some interviews
and so some statements might not have been recorded.  Tr. at 377-8.
Second, Dr. Hamilton testified that portions of the statements which
were recorded were difficult or impossible to transcribe, although "the
transcript is as good a representation of the tape as one could get."
Tr. at 387-8.  The tape itself was provided for the record and upon
review, it is evident that there were inaudible portions although the
quoted sections appeared to be transcribed accurately.  Exs. 29 and 29A.
It is therefore impossible to be certain that the statements by Dr.
Sharma represented a full picture of his explanations to the Hamilton
Committee.  Third, there is uncontradicted evidence in the record that
Dr. Sharma's request to have counsel with him in the interview was
denied and that he was not permitted to review the transcript to "make
any clarifications, corrections or comments."  Ex. R-2, Part II, at 2-3;
Ex. H-31, at 312-5.

27.   The page length reference is more consistent with the revised
grant application, Exhibit H-35, which was 47 pages.  The original
application was 52 pages.  Ex. H-2.

28.   ORI was particularly swayed in its assessment of Dr. Sharma's
credibility by its review of Dr. Sharma's interview by the Hamilton
Committee.  Tr. at 689-90 (Dr. Lyle Bivens of ORI); see also ORI Final
Report at 49-50.  ORI viewed the "crux" of that "presentation" as the
claim by Dr. Sharma to have "had a reasonable basis for anticipating the
à2GC amino acid sequence data," while ORI contended he could not have
expected that data because Dr. Marshak of Cold Spring Harbor had already
told him that no further work could be done on it.  ORI Final Report at
30.  Dr. Sharma presented documentation about his collaboration with Dr.
Marshak, but later clarified that he was "aggressively pursuing
obtaining" à2 sequence (from Dr. Marshak as well as at least two other
laboratories with which he documented earlier collaborations) "before
the other laboratory [Lefkowitz] reported cloning of this receptor."
Ex. H-23, at 2 (Letter from Dr. Sharma to OSIR).  However, he stated
that "[a]fter this receptor [à2A] was cloned our enthusiasm for
sequencing à2GC subsided.  We were then only interested in obtaining its
sequence sometime in the future for the sole purpose of identifying one
of our isolated clones as the à2GC."  Id.  This position is consistent
with the weight of evidence at the hearing that comparison of potential
clones obtained with the à2A probe with à2GC sequence data once
available was a reasonable research strategy.  See, e.g., Tr. at 821,
954-5.  Dr. Sharma could not have expected imminent receipt of sequence
data from Dr. Marshak during the time frame when the application was
prepared and submitted, because Dr. Marshak had informed Dr. Sharma or
his staff some time in 1988 that the samples of à2GC had not yielded
useful data, and no new samples had been sent.  Tr. at 187-90.  However,
ORI did not prove that Dr. Marshak made clear to Dr. Sharma before 1990
that Dr. Marshak would do no further work on the project, although Dr.
Marshak did apparently communicate on several occasions that his
participation was becoming more "difficult."  Tr. at 191.  Only in May
of 1990 did Dr. Marshak return the samples and end the collaboration.
Ex. H-11, at 1.  This chronology further supports Dr. Sharma's position
that he was not expecting the data for this application but continued to
anticipate that further efforts might yield sequence data for future
comparison with the clones he had obtained.  Dr. Sharma testified that
the reason he brought the Cold Harbor correspondence to the committee
was that he had gotten the "impression that there was some doubt that I
had not started the work on the sequence . . . ."  Tr. at 822.  Some of
his inexplicable comments to the Hamilton Committee may also have
reflected this additional layer of confusion about whether he was being
asked to prove that he really had made some independent efforts to get
à2GC before deciding to use Lefkowitz' probes after the à2A data was
published.

29.   ORI argued that Dr. Hamilton's opinion that Dr. Sharma did offer
an anticipatory writing explanation was based on a review of "all the
tapes of the [Hamilton] . . . Committee, not just the transcribed one."
ORI Post-Hearing Br. at 36.  We cannot give additional weight to his
statements on that account, since ORI did not submit the additional
tapes into evidence or offer any explanation for failing to do so.  In
any case, despite ORI's characterization of Dr. Hamilton's testimony at
the hearing as having "clearly remembered Dr. Sharma's conflicting
explanations," as quoted in the text, Dr. Hamilton still appeared to be
uncertain even at the hearing whether or not Dr. Sharma told his
Committee that he did intend to insert à2GC in the hope of obtaining
data in time to justify it.  See ORI Post-Hearing Br. at 36.

30.   See Tr. at 41 and 44, identifying the application which Dr.
Chalberg saw first as Exhibit H-35.

31.   Although Dr. Chalberg wrote basic protocols of the molecular
biology methodologies in the laboratory, which were in the computer
database to be inserted into grant applications or writings by Dr.
Sharma, Dr. Chalberg repudiated any claim to be an author of any of
these grants.  Ex. H-17, at 1.

32.   Dr. Bannerjee's recollection of whether Dr. Sharma stated at the
April 20th meeting that he wrote à2GC in the disputed sentence because
he hoped to get the data in time to justify it was considerably less
certain than Dr. Ferrario's.  At the hearing, Dr. Bannerjee testified
that he "think[s] what happened" was that he was called in to Dr.
Ferrario's office to hear Dr. Sharma's explanation.  Tr. at 642.  He
reported that Dr. Sharma said that the "real thrust of the research" in
Exhibit H-35 was "something else" and that Dr. Sharma said he "did not
know what this problem is all about and he has to look into this matter
more."  Id.  All Dr. Bannerjee remembered of the discussion about the
à2GC receptor was that Dr. Sharma said he did not have the sequence but
that in any case in the revised application he had switched to using the
cDNA clone in rat brain research, and "something was left over I guess
in the background section."  Id. at 643.  When Dr. Bannerjee was asked
if he had "any doubt in [his] mind sitting there today that the Alpha
2GC was discussed with Dr. Sharma" in the April 20th meeting, all he
responded was that he "think[s] it was discussed in that meeting."  Id.
at 644.  Dr. Bannerjee made absolutely no reference to the anticipation
explanation which Dr. Ferrario understood Dr. Sharma to have been
offering at that meeting.  Even more striking is a question by Dr.
Douglass to Dr. Sharma during the latter's second OSIR interview, in
which Dr. Douglass quotes at length from Dr. Bannerjee's OSIR interview
(which was not submitted for the record).  There, Dr. Bannerjee was
directly asked if Dr. Sharma said he submitted the application hoping to
get the data before the review.  Dr. Bannerjee was quoted as responding
that "he did not say anything about that in our first meeting because
that meeting did not last more than ten minutes, you know, in Ferrario's
room."  Ex. H-31, at 361.  Thus, Dr. Bannerjee does not corroborate Dr.
Ferrario's account.

33.   Mr. Murray (the administrator who attended the meeting) added
little to the question of what explanation Dr. Sharma may have offered
then, if indeed he, as a non-scientist, clearly understood the precise
charges.  Mr. Murray stated only that Dr. Sharma admitted he did not
have the sequence data, though he understood that Dr. Sharma had hoped
to.  Tr. at 257.  Mr. Murray's contemporaneous memorandum had no
reference to anticipation of data, but simply said that "Dr. Sharma
readily admitted that Dr. Chalberg's claims were accurate."  Ex. H-9.
While he thereby agrees with Dr. Ferrario that a meeting occurred in
which Dr. Sharma was told of charges of scientific misconduct relating
to an à2GC grant and admitted that any claim to have à2GC sequence data
was erroneous, he does not corroborate Dr. Ferrario's perception that
Dr. Sharma said he wrote the sentence because he thought would have the
data before the grant was reviewed.  Cf. Ex. H-4, at 2 (Dr. Ferrario's
memorandum).

34.   An additional point which must be noted is that, according to the
Healy Committee report, Dr. Healy, then the head of the Research
Institute, had asked Dr. Ferrario about a week before these charges were
raised by Dr. Chalberg whether "there was any evidence of poor
performance or scientific misconduct" by Dr. Sharma to explain the
"recent discontent in his section."  Ex. H-5, at 1.  Dr. Ferrario said
there was not, and mentioned that Drs. Chalberg and Jaiswal had not
brought any such complaints to him.  Id. at 1-2.  A little over a week
thereafter, Dr. Ferrario gave copies of Dr. Sharma's applications to Dr.
Chalberg to review, and allegations of misconduct resulted.  In light of
such coincidences, and the clear evidence of conflict within the
laboratory and the department, we cannot accept as proven ORI's claims
that Drs. Ferrario and Chalberg acted reluctantly and against their own
interests in attacking Dr. Sharma at the time these charges arose.  Cf.
ORI Post-Hearing Br. at 39-40.

35.   Further, while ORI stated that Dr. Sharma's complaints against Dr.
Ferrario were found to be groundless by the Hamilton Committee, ORI
cited nothing in the record to that effect.  ORI Post-Hearing Br. at 40.
The Hamilton Committee was instructed to review the complaints relating
to Dr. Ferrario, but did not include any response in its report.  Exs.
H-6, H-7.  The Healy Committee report stated that "it was unacceptable
practice" for the department chairman to hire staff from a senior
investigator's laboratory without his consent to get "special technology
that that senior investigator offers."  Ex. H-5, at 5.  This conclusion
suggests that Dr. Sharma's suspicions concerning the motives of his
accusers may have had some basis in reality.  Cf. id. at 4; ORI
Post-Hearing Br. at 40.

36.   The other statements, with the problematic clauses highlighted,
were as follows:

 Two oligonucleotide-39-mer probes, based on rat codon usage
 (98), corresponding to the amino acid sequences of two
 internally cleaved peptides, were synthesized and used to screen
 the unamplified þ ZAP (Statagene Cloning Systems) rat
 adrenocortical carcinoma cDNA expression library of two million
 phage.

 Two oligonucleotide-39-mer probes, based on rat codon usage
 (98), corresponding to the amino acid sequences of two
 internally-cleaved peptides, were used to screen the phage
 libraries.

Exhibit H-2, at 27 and 32 (emphasis added).

37.   ORI also relied on Dr. Sharma's failure to notify NIH in relation
to the claim that such negligent conduct showed "disregard for ensuring
the accuracy" of grant applications and constituted scientific
misconduct.  ORI Post-Hearing Br. at 30.  This claim is addressed above.

38.   This statement is considerably less forceful than the paraphrase
of it by ORI to read such an error "warranted immediate notification to
NIH."  ORI Post-Hearing Br. at 30, n.80 (citing Tr. at 584-5).  In fact,
the actual statement by Dr. Douglass does not clearly set out an
accepted standard for when to notify NIH, but rather indicates that
researchers use discretion in evaluating whether an error is
significant.  However, Dr. Sharma himself agreed that he would have a
duty to inform NIH if a significant error existed in an application
which was before NIH for possible funding.  Tr. at 847-8.

39.   Both Mr. Murray and Dr. Ferrario indicated at the hearing that Dr.
Sharma was not summoned by Mr. Murray, but rather was called by Dr.
Ferrario's secretary.  Tr. at 297; Tr. at 417-8.

40.   Dr. Bannerjee's interview with ORI supported the description of
Dr. Sharma's participation in the meeting as brief, five or ten minutes.
Ex. H-31, at 361. Mr. Murray stated that Dr. Sharma's participation
lasted not more than 30 minutes.  Tr. at 296.  Dr. Ferrario had no idea
how long the meeting lasted.  Tr. at 421.

41.   "If you were trying to use rat to go to human, you'd have to look
up a human."  Tr. at 571.  Thus, Dr. Sharma would have had to "look up"
rat data to use human sequence for a probe "to go to" rat cells.

42.   ORI argued that Dr. Sharma intended to mislead and was prevented
only because Dr. Chalberg wrote a portion of the application (on page 35
of Exhibit H-3) which stated that probes would be synthesized "as soon
as sufficient sequence data" became available.  ORI Post-Hearing Br. at
26.  This argument is without merit, since Dr. Chalberg made no claim to
have written the numerous other contradictory statements in the
application, including those cited by the reviewers.  Ex.